Paralysis by Inertia “Like” Habit in Social Networking Services: Tendency to Answer Loyalty Questions in Marketing Surveys

Abstract

Factors affecting survey quality, such as questionnaire length, screen design, incentives for respondents, and answering devices have been extensively researched before. This study introduces a new dimension to the factors of response bias’ factors and focuses on the possibility that Like habits and motives in social networking services (SNS) may lower the psychological barrier to favorable reactions for brand loyalty (e.g., recommendation intention). Propensity score was applied to the results of Japanese smartphone users’ online survey, and the effect of Like habits on response tendencies was verified. Results show that compared to the group that did not use SNS, there were no differences with those who used true Likes, and a group of inertia Like has a high score—becoming clear that the psychological barriers differ depending on Like motive. If the SNS usage status of survey respondents fluctuates, the results may be biased and mistakes could occur in decision-making.

Keywords

Like button motive propensity score matching recommendation smartphone

Building long-term customer loyalty contributes to efficient corporate profits. This is because customers with high loyalty have a high repurchase rate (Oliver, 1999) and often recommend brands to acquaintances (Aaker & Joachimsthaler, 2000). Hence, the number of highloyalty customers greatly influences the stability and growth potential of corporate management. Various indices are used to measure loyalty. In particular, being confident enough to recommend it to others is a sign of commitment to the brand (Aaker, 1991). Psychological barriers are higher for one’s willingness to encourage others to recommend compared with the satisfaction/preferences generated by individual emotions. Moving beyond such barriers and having recommendation intentions is a strong indication of interest in a brand. Based on this concept, recommendation intentions have been generally used in marketing surveys that measure loyalty in both industry and research.

However, with the rise of social networking services (SNS) and the creation of the “Like” button, the act of recommendation has become quite common. Likes, shares, and comments are common forms of customer engagement on social media platforms (Kabadayi & Price, 2014). Among them, the Like button is the simplest and easiest way to engage with one click (Gavilanes et al., 2018; C. Kim & Yang, 2017; Liu et al., 2017), and more people tend to Like posts than comment on them (Gerodimos & Justinussen, 2015). Owing to the simplicity of the action, the motives behind liking social media content are ambiguous. Consumers develop a habit of pressing the Like button unconsciously or to maintain social relationships, regardless of whether they feel the content of the post is good (Chin et al., 2015; Egebark & Ekström, 2018; Eranti & Lonkila, 2015; Hayes et al., 2016; Lee et al., 2016; Leiner et al., 2018; Levordashka et al., 2021; Sumner et al., 2018).

Thus, in relation to motives, Likes are classified into two types. The first is a “true Like,” when consumers feel good and want to recommend a post’s content to their acquaintances. The other is an “inertia Like,” such as liking a post to maintain relationships with acquaintances, returning Likes from others, and unconsciously liking posts while browsing social media. In particular, the latter habit has taken root in consumers’ daily lives.

The effects of Likes on consumer behavior have been discussed in many previous studies (Beukeboom et al., 2015; Ding et al., 2017; Egebark & Ekström, 2018; Sherman et al., 2016; Sorensen et al., 2017; Thai & Wang, 2020). Since it is difficult to judge individuals’ motives behind or feelings regarding Likes from data, previous studies have mainly focused on number of Likes (Ji et al., 2017). Therefore, research on inertia Likes is overwhelmingly insufficient. Furthermore, this study focused on the impact of inertia Likes on response trends in marketing research, which was surprisingly uncontroversial. In the present study, it was assumed that it is possible inertia Like habits in SNS could reduce the psychological barriers to favorable reactions in marketing surveys. For verification, an online survey was conducted to examine consumer attributes, SNS usage, and loyalty indices for smartphone brands, and the influence of SNS Like habits/motives on response tendencies was subsequently examined. In this study, the following four hypotheses were verified; H1: Compared to those who do not use SNS, those who do use SNS but do not have a Like habit will show no differences in scores for loyalty indices on surveys; H2: Compared to those who do not use SNS, those who have become accustomed to providing true Likes on SNS will have higher scores for loyalty indices on surveys; H3: Compared to those who do not use SNS, those who have become accustomed to providing inertia Likes on SNS will have higher scores for loyalty indices on surveys; H4:The effect size of inertia Likes, as in H 3, will be larger than the effect size of true Likes in H2. If it is evident that Like habits affect survey responses, companies should consider that SNS usage in a sample is constant when conducting regular marketing surveys. Accordingly, this study provides useful suggestions for both industrial and academic research.

Literature Review and Hypotheses

Like’s Influence on Consumer Behavior

As shown in the consumer behavior model AISAS (Attention → Interest → Search → Action → Share; Kono, 2009), it is common for consumers to voluntarily search for information before the purchase and post reviews after it. When searching for information, consumers refer to word-of-mouth as well as official company information. This is because word-of-mouth expresses what the consumer, who has experienced the product/service, feels without any commercial intention. SNS users like the opinions of others (Bastos, 2015). There are around 4.62 billion SNS users in 2022, which is 3.1 times that of 2012, and the mean usage time of SNS is 2 hr 27 min per day (Kemp, 2022). Additionally, the size of the SNS market is expected to continue to grow.

The “Like” button is one of the factors that supports the growth of SNS. The Like button was first introduced by the video site Vimeo in 2005, but it became widespread globally when it was adopted by Facebook in 2009 (Moffat, 2019). At the beginning of the development of Facebook, adding an “Awesome” button was considered. Facebook had concerns that users would find the button annoying (Khrais, 2018); however, the ability to easily express support for connected acquaintances and recommend content to others resulted in an environment that motivates consumers to make posts.

Exchanging opinions about products and services with friends on SNS greatly influences subsequent purchase behavior (X. Wang et al., 2012). In particular, the Like button has a significant impact on the recipient of information. Social endorsement via Likes tends to garner trust because it is voluntary rather than driven by commercial intentions (Thai & Wang, 2020). For example, impressions of posted photos have been shown to vary depending on the Like count (Sherman et al., 2016). Likes also have a huge impact on businesses (Lipsman et al., 2012; Trattner & Kappe, 2013), as companies are more often concerned about the number of Likes than that of fans (Naylor et al., 2012). BestBuy.com and more than 350,000 other websites installed a Like button within a few months of Facebook’s appearance (Gelles, 2010). Clicking “Like” on a brand page on Facebook contributes to brand attitude, such as brand love, purchase intention, and recommendation intention (Beukeboom et al., 2015; Kudeshia et al., 2016; Pöyry et al., 2013). In fact, a 1% increase in a movie’s prerelease Like count has been shown to increase box office revenue in the first week by approximately 0.2% (Ding et al., 2017). Since such an effect has been confirmed, research on activities that increase or have increased the number or effect of Likes has also been conducted (López et al., 2017; VanMeter et al., 2018). As indicated by these previous findings, companies should actively manage consumers experiences in the SNS brand community to encourage dialog and endorsements between consumers (X. W. Wang et al., 2019).

In addition, Like data are used in academic research as a source of information that expresses user preferences. Moreover, sensitive personal attributes, such as ethnicity and political views, can be inferred (Kosinski et al., 2013). It was also reported that computerbased judgments of people’s personalities based on Like data are more accurate and valid than judgments made by their close others or acquaintances (e.g., friends, family, spouse, colleagues [Youyou et al., 2015]). In the 2008 US presidential election, Like count also revealed that photos with Mrs. Michelle Obama and her daughters, as well as Mr. Obama, had an even more engaging effect, compared to Mr. Obama’s photo (Gerodimos & Justinussen, 2015). Thus, as data generated from the Like button are useful, such data are also called the “like economy,” and are actively used in both industry and research (Gerlitz & Helmond, 2013).

Many reports have shown that Likes represent a positive attitude toward brands and brand content. Therefore, considering that consumers habitually repeat favorable behaviors, three hypotheses were derived. H1 was derived based on the assumption that there would be no differences between both those who do and do not use SNS without a Like habit in favorable response tendencies (e.g., preference, satisfaction, repurchase intention, recommendation intention). For H2 and H3, it was assumed that for SNS users, a Like habit would lead to lower psychological barriers for favorable reactions, compared to those who do not use SNS, regardless of motives.

H1: Compared to those who do not use SNS, those who do use SNS but do not have a Like habit will show no differences in scores for loyalty indices on surveys.

H2: Compared to those who do not use SNS, those who have become accustomed to providing true Likes on SNS will have higher scores for loyalty indices on surveys.

H3: Compared to those who do not use SNS, those who have become accustomed to providing inertia Likes on SNS will have higher scores for loyalty indices on surveys.

Inertia Like Habituation on SNS

Consumer inertia behavior has been emphasized in academic research, with the most active discussions on spurious loyalty. Loyalty cannot be judged only by the number of purchases. Factors like inertial continuous and forced continuous purchases should also be considered. It also includes the case where the calculation continues in consideration of the loss that occurs when the relationship with the brand ends. In other words, loyalty cannot be judged without looking at behavioral indices as well as attitude indices (Assael, 1987; Cushing & Douglas-Tate, 1985; Dick & Basu, 1994; Gilliland & Bello, 2002).

Inertial continuous purchase behavior has been reported in studies in various industries. In online shopping, consumer inertia, rather than satisfaction was found to contribute to repeat purchases (Kuo et al., 2013). In the residential electricity market, consumers rarely search for alternatives, and inertia is likely to occur (Hortaçsu et al., 2017). Inertia has shown a stronger influence in healthcare than in other service sectors (Gonçalves et al., 2020). Factors of inertia in purchase behavior include product/service similarity, financial/time costs in brand switching, learning costs, and habituation (B. Kim & Kang, 2016; White & Yanamandram, 2004).

Inertia is also important in relation to Like behavior on SNS. Along with the influence of Likes as discussed above, it should be noted that Like motivation is not a simple positive evaluation (Leiner et al., 2018). Aside from the positive motivation for users to genuinely express that they like SNS content (Kudeshia et al., 2016; Munar & Jacobsen, 2014; Swani et al., 2013) or provide their opinions (won J. W. Kim, 2014), Likes can be used passively, such as for maintaining social ties with acquaintances (Chin et al., 2015; Levordashka et al., 2021; Sumner et al., 2018), pleasing others (Lee et al., 2016), and rewarding those who Liked the user’s own posts (Eranti & Lonkila, 2015). Alternatively, Likes act as a form of aimless automated communication (Hayes et al., 2016). In fact, an unconscious tendency to like posts regardless of the content has been clarified. Users are less likely to prefer posts that only one stranger has liked, but tend to Like posts that three strangers have liked (Egebark & Ekström, 2018). Therefore, consumers will develop a habit of succumbing to Like inertia, regardless of whether content is good or bad. The reason why users can press the Like button unconsciously or easily is that liking is the simplest action that can be performed on SNS, compared to commenting and sharing (Gavilanes et al., 2018; Gerodimos & Justinussen, 2015; C. Kim & Yang, 2017; Liu et al., 2017). In this context, ease can include both the amount of activity and the psychological burden that comes with users’ concerns that commenting poorly or sharing lowquality content with their comments added could damage their status and credibility (Thompson et al., 2020; Yoon et al., 2018).

However, academic research also indicates that Likes are limited in their ability to capture users’ true intentions. Reacting to SNS content by providing a comment shows the strength of a relationship more than making one click to Like content. Moreover, inperson and telephone contact are more effective than all forms of SNS content interaction (Burke & Kraut, 2014). Consumers are less likely to buy a brand when they learn that their friends like the brand through Likes on social media than when they learn through offline means that a friend genuinely likes a brand (John et al., 2017).

Thus, inertia Likes are used without the act of recommendation maintaining its significance. When inertia Likes are habitual, attention is paid to the possibility that these Likes increasingly lower psychological barriers to favorable reactions when compared to true Likes. Therefore, a fourth hypothesis was as developed as follows:

H4:The effect size of inertia Likes, as in H3, will be larger than the effect size of true Likes in H2.

Since it is difficult to judge consumers’ motivations and feelings from data on Likes, previous studies have mainly focused on the number of Likes received (Ji et al., 2017). Considering the lack of method for expression, Facebook has developed a reaction function that can express Love, Care, Haha, Wow, Sad, and Angry as an extension of Like (Facebook, n.d.). However, consumers still enjoy the simple action of clicking the Like button, including users of various SNS such as Twitter, YouTube, and TikTok. Therefore, the impact of consumers’ underlying motives should be carefully observed. As shown in Table 1, and to the best of my knowledge, this study is the first to focus on the effects of inertia Likes on response trends in marketing research.

Table 1.

Previous Research on the Effects of Likes When Using SNS.

Direction of effect	Paper	Consequences	Motivation for “Like”
Positive	Pöyry et al. (2013)	Participation in the brand community by Likes enhances purchase intention and referral intention.	True Likes such as truly enjoyable
	Beukeboom et al. (2015)	Consumers who continue to like a brand’s Facebook’s page are more likely to have purchase intention and recommendation intention.	—
	Kudeshia et al. (2016)	Likes on the fan page of a brand contributes to the brand love, and enhances word-of-mouth posts and purchase intention.	Mainly true Likes such as · find products to be unique · To know more about it
	Sherman et al. (2016)	Participants were more likely to like photos with many likes than photos with few likes.	−
	Ding et al. (2017)	A 1% increase in a movie’s pre-release Like count has been shown to increase box office revenue in the first week by approximately 0.2%.	−
	Sorensen et al. (2017)	Interactions (likes, clicks, shares, and retweets) in SNS also have the effect of promoting co-creation of value.	−
	Egebark and Ekström (2018)	Users are less likely to prefer posts that one stranger Liked, but tend to express positive support for posts that three strangers Liked.	−
	Thai and Wang (2020)	Liking behaviors result in enhanced brand attitude, engagement, brand identification, and purchase intention.	−
Negative	Burke and Kraut (2014)	Reactions by comments show the strength of the connection of the relationship more than Likes that can be done by a one-click action.	−
	John et al. (2017)	Consumers are less likely to buy the brand when they know that their friends like the brand through Likes on social media than when they learn that a friend genuinely likes the brand in the offline sense.	−

Research Method

Survey Method

In this study, an online survey was conducted in Japan from January 15 to 21, 2021. Respondents were unaware of the purpose of this study, which was to investigate the relationship between behavior on SNS and response tendencies in marketing surveys. Smartphones were selected as a product close to SNS. The title of the survey was displayed as “Survey on Smartphones.” The two conditions for the respondents were that their ages were between 20 and 60 and they possessed a smartphone. The sample size was calculated using the following: 5 generations ×’ 2 genders = 10 categories. Each category had 60 people: hence, a total of 600 people. The survey was distributed to a panel owned by a research company in Japan to achieve the target sample size.

An online survey points to minimization of an effort called “Satisficing” that can reduce the reliability of responses (Krosnick, 1991). Therefore, to improve reliability, it is effective to shorten the number of questions in the survey as much as possible and reduce the burden on the respondents. Based on this, the following 15 items were defined: (1) gender, (2) age, (3) area of residence, (4) annual household income, (5) frequency of Read, (6) frequency of Like, (7) frequency of Post, (8) motive of Like, (9) owned smartphone brand, (10) purchase elapsed time, (11) purchase emphasis point, (12) preference, (13) satisfaction, (14) repurchase intention, and (15) recommendation intention. The distribution of respondent attributes for items (1)−(4) is shown in Table 2. The information on the device was automatically acquired by the survey system. Device data were collected because the answering tendency differs depending on the device (Lugtig & Toepoel, 2016; Mavletova, 2013), so the influence was taken into consideration. Items (5)−(7) were asked in relation to Twitter, Facebook, Instagram, and TikTok, which are the most commonly used SNS in Japan. The choices were as follows: “not at all,”“less than once a week,”“two to three times a week,” four to six times a week,” and “every day.” Like and Post were controlled, so that options higher than the frequency of Read could not be selected. In (6), there were restrictions on the targets of Like (e.g. acquaintances, celebrities, brands).

Table 2.

Distribution of Respondent Attributes.

Item	Breakdown	Number of respondents	Item	Breakdown	Number of respondents
Device	Smartphone	348	Income	2–4 million yen	61
Device	Personal computer	252		4–6 million yen	144
Gender	Male	300		6–8 million yen	160
Gender	Female	300		8–10 million yen	101
Age	20 s	120		10 million–15 million yen	65
	30 s	120		15 million yen or more	69
	40 s	120	Purchase elapsed time	Less than a year	154
	50 s	120		1 to 2 years	197
Area	Hokkaido	31		2 to 3 years	108
	Tohoku	27		More than 3 years	141
	Kanto	251	Brand	Apple	263
	Chubu	78		Fujitsu	27
	Kinki	128		Samsung	44
	Chugoku	37		Sharp	99
	Shikoku	13		Sony	73
	Kyusyu	35		Others	94

As shown in Table 3, Twitter was the most used in all cases. The results of extracting Read, Like, and Post of the most frequently used SNS of each respondent are at the bottom of Table 3. Item (8) presented the following four motives for Likes: (a) Like feeling really good, (b) Like telling the fact that you saw it, (c) Like in return, (d) Somehow Like. Here, (a) is a true Like, and (b)–(d) are inertia Likes. Then, numerical answers were requested, so that the ratio of each motive would be 100% in total. It was set such that only answers with a total of 100% would be accepted. As shown in Figure 1, the answers were concentrated in a nice round number because the answers were free input, but real Likes were distributed in a well–balanced manner from 0% to 100%. Inertia Likes were most often 0%, but were continuously distributed, up to approximately 50%. Items (9)−(15) were questions about smartphone brands. In (9), Apple, Fujitsu, Samsung, Sharp, and Sony, which were the top five smartphone brands in Japan in 2020 (MM Research, 2021 (Sudarshan, 2021)), were presented as options. The distributions of (9) and (10) are shown in Table 2. In (11), design, performance, price, quality, reputation, usability, and others were presented. Questions (12)−(15) were loyalty indices, which were scored using five scales (e.g., 1: very unsatisfied, 7: very satisfied).

Table 3.

SNS Usage Status.

	SNS	Frequency					Total	Active ratio (%)
	SNS	Not at all	Less than once a week	Two to three times a week	Four to six times a week	Every day	Total	Active ratio (%)
Read	Twitter	303	48	30	38	181	600	49.5
	Facebook	377	90	34	24	75	600	37.2
	Instagram	364	54	27	33	122	600	39.3
	TikTok	544	24	8	7	17	600	9.3
Like	Twitter	426	68	36	20	50	600	29.0
	Facebook	474	68	25	13	20	600	21.0
	Instagram	450	64	25	21	40	600	25.0
	TikTok	567	15	4	5	9	600	5.5
Post	Twitter	455	72	25	15	33	600	24.2
	Facebook	512	61	9	7	11	600	14.7
	Instagram	489	75	14	6	16	600	18.5
	TikTok	579	8	4	3	6	600	3.5
Total	Read	214	54	38	43	251	600	64.3
	Like	337	101	47	36	79	600	43.8
	Post	389	124	27	18	42	600	35.2

Figure 1.

Distribution of Like motives.

As shown in Table 4, when looking at means by brand, Apple had the highest score for all indices. By index, recommendation intention had the lowest mean. As mentioned above, these results indicated that the psychological barrier was higher for the intention to recommend to others than the preference, satisfaction, and repurchase intention that are completed by individual emotions.

Table 4.

Means of Loyalty Indices for Each Smartphone Brand.

	Apple	Fujitsu	Samsung	Sharp	Sony	Others	All brands
Preference	3.875	3.444	3.477	3.586	3.726	3.191	3.653
Satisfaction	3.871	3.519	3.659	3.535	3.493	3.277	3.645
Repurchase	3.962	3.407	3.341	3.293	3.507	2.840	3.550
Recommendation	3.567	3.037	3.136	2.990	3.164	2.798	3.247

Here, groups were defined to test hypotheses. As shown in Table 5, from the results of (5)−(6), 214 people who did not Read or Like were Group 1 (No SNS use), and 123 people who Read but do not Like comprised Group 2 (No Like use). The remaining 263 were categorized as Like motives. In (8), the flag of true Like was set when (a) was 50% or more, and the flag of inertia Like was set to 1 when (b)–(d) was 50% or more in total. If both were exactly 50%, both flags were set to 1. As shown in Table 6, 109 people with a high percentage of true Likes were classified as Group 3 (true Like), and 116 people with a high percentage of inertia Likes were classified as Group 4 (inertia Like). There were 38 respondents with an equal proportion of both true Likes and inertia Likes, and they were excluded due to the small sample size. Table 7 shows the definition of each group and a list of the number of respondents per group.

Table 5.

Group Classification Based on Read and Like Frequency (Group 1, Group 2).

			Like					Total
Frequency			1	2	3	4	5	Total
Read	1	Not at all	214	0	0	0	0	214
	2	Less than once a week	32	22	0	0	0	54
	3	Two to three times a week	19	10	9	0	0	38
	4	Four to six times a week	20	12	6	5	0	43
	5	Every day	52	57	32	31	79	251
Total			337	101	47	36	79	600

Table 6.

Group Classification Based on Like Motive (Group 3, Group 4).

		Inertia like		Total
		No	Yes	Total
True Like	No	0	116	116
True Like	Yes	109	38	147
Total		109	154	263

Table 7.

Definitions for Each Group.

Group	Definition	Number of respondents
Group 1: No SNS use	People who do not use SNS	214
Group 2: No Like use	People who read but do not Like on SNS	123
Group 3: True Like	People who read and Like on SNS and have a true Like ratio of 50% or more of all Likes	109
Group 4: Inertia Like	People who read and Like on SNS and have an inertia Like ratio of 50% or more of all Likes	116
Not applicable	People who read and Like on SNS, and the ratios of true Likes and inertia Likes out of all Likes are exactly 50% each	38
Total		600

Verification Method

If randomization is possible, it is ideal to conduct a randomized controlled trial, which is the most reliable scientific basis (Torgerson & Torgerson, 2008). However, it is difficult to ask randomly selected people to specify their frequency of SNS use and to implement it for a certain period. In addition, liking content as a habit is more appropriate rather than forcibly pressing the Like button.

The propensity score proposed by Rosenbaum and Rubin (1983) is a typical method for estimating causal effects when randomization is difficult. The covariates are adjusted by aggregating multiple covariates into one variable (i.e., the propensity score). This is because, in each group, the characteristics of consumers were biased. For example, as shown in Table 8, when comparing the respondent attributes of each group, Group 1, which did not use SNS, had a higher age bracket than other groups. The tendency of the younger generation to have a higher SNS usage rate is common worldwide (Schumacher & Kent, 2020). In addition, looking at the distribution of brands, there was a large gap in Apple usage rate (Brand Apple) between Group 1 and Group 4. Accordingly, it would have been meaningless to compare the response tendencies of loyalty indices. In the present study, Group 4 had the highest scores for all loyalty indices, but this was affected by various factors and no conclusion could be drawn.

Table 8.

Variable List and Means for Each Group.

No	Variable	Description	Breakdown	Group
No	Variable	Description	Breakdown	1	2	3	4
1	Preference	Preference	5 scales	3.561	3.675	3.624	3.819
2	Satisfaction	Satisfaction	5 scales	3.523	3.634	3.670	3.819
3	Repurchase	Repurchase intention	5 scales	3.383	3.545	3.560	3.810
4	Recommendation	Recommendation intention	5 scales	3.042	3.276	3.284	3.500
5	PC	Dummy of response device	PC	0.486	0.350	0.376	0.440
6	Male	Dummy of gender	Male	0.481	0.480	0.523	0.517
7	Age	Age	1: 20s, …,5:60s	3.565	2.919	2.523	2.647
8	Area_01_Hokkaido	Dummy of residential areas	Hokkaido	0.070	0.081	0.037	0.009
9	Area_02_Tohoku		Tohoku	0.033	0.033	0.055	0.078
10	Area_03_Kanto		Kanto	0.453	0.407	0.358	0.431
11	Area_04_Chubu		Chubu	0.107	0.114	0.165	0.155
12	Area_05_Kinki		Kinki	0.192	0.220	0.220	0.241
13	Area_06_Chugoku		Chugoku	0.061	0.065	0.064	0.052
14	Area_07_Shikoku		Shikoku	0.019	0.033	0.018	0.009
15	Area_08_Kyusyu		Kyusyu	0.065	0.049	0.083	0.026
16	Income	Household income (million yen)	1: <4, 2: 4–6;3: 6–8, 4: 8–10,5: 10–15,6: >15	3.294	3.285	3.303	3.284
17	Purchase_Time	Purchase elapsed time	1:< a year, …, 4: > 3 years	2.421	2.276	2.523	2.397
18	Reason_Design	Dummy of purchase emphasis point	Design	0.061	0.130	0.110	0.060
19	Reason_Performance		Performance	0.215	0.146	0.248	0.267
20	Reason_Price		Price	0.257	0.268	0.257	0.172
21	Reason_Quality		Quality	0.084	0.057	0.064	0.086
22	Reason_Reputation		Reputation	0.023	0.016	0.018	0.034
23	Reason_Usability		Usability	0.252	0.268	0.220	0.233
24	Reason_Others		Others	0.107	0.114	0.083	0.147
25	Brand_Apple	Brand owned	Apple	0.341	0.407	0.505	0.543
26	Brand_Fujitsu		Fujitsu	0.037	0.065	0.055	0.034
27	Brand_Samsung		Samsung	0.093	0.089	0.037	0.060
28	Brand_Sharp		Sharp	0.187	0.138	0.183	0.129
29	Brand_Sony		Sony	0.168	0.130	0.073	0.103
30	Brand_Others		Others	0.173	0.171	0.147	0.129
	Sample size			214	123	109	116

Therefore, causal effects were estimated by matching respondents with similar propensity scores and homogenizing both groups. In this study, propensity score matching was performed for each of the two paired groups (Groups 1 and 2, Groups 1 and 3, and Groups 1 and 4). In the case of three or more groups, the method of homogenization by generalized propensity score matching was not adopted, owing to the insufficient sample size. When the true value of the propensity score of each respondent is unknown, as was the case in the present study, the method of estimating from the data using a logistic regression model is common. Numbers 5 to 30 shown in Table 8 were set as the explanatory variables. Numbers 8, 24, and 30 were set as the criteria of the dummy variable and not used in the model. For the objective variable, 0/1 data indicating either Group 1 (control group) or Groups 2–4 were input. Thus, the propensity score represented the probability that each respondent would belong to Groups 2–4. Since there were many explanatory variables used when building the model, variable selection was performed using the stepwise method.

The difference in the distribution of loyalty indices was verified by Fisher’s exact test for the two groups matched using propensity scores. The chi-square test was not used because the cells in the contingency table contained numbers less than 10. The null hypothesis was that there would be no difference in the distribution of loyalty indices between the two groups. The applied significance level was 5%, and the null hypothesis was rejected when the p-value was less than .05. An alternative hypothesis was adopted: there would be a difference in the distribution of loyalty indices between the two groups. This test verified H1−H3. H4 was compared with Cramer’s V, which represents the effect size. Generally, it is judged as follows: 0.1 is a small effect size, 0.3 is medium, and 0.5 is large (Grant et al., 2012; Khalilzadeh & Tasci, 2017).

To carry out the test strictly, the validity of the sample size was confirmed. The appropriateness of the sample size was confirmed using a power analysis. A total of 132.614 people were calculated based on the following conditions: significance level, 5%; power, 80%; effect size, medium 0.3 (Cohen, 1992); and the degree of freedom, 4 (2 groups, matrix of loyalty indices in 5 scales). It was confirmed that the total number of people in both groups needed to be 133 or more. If the sample size was excessively large, there would be a concern that even a meaningless difference may be detected. The effect size was 1,193.529 when calculated with a small effect size (0.1) and 298.382 when calculated with an effect size of 0.2. Since there was a concern that excessive detection would occur if these values were exceeded, care was taken not to exceed 300 people per group.

It should be noted that statistical analysis was conducted using the software program R. The stepwise method used was the stepAIC function of the MASS package. Propensity score matching was conducted using the Match function of the Matching package, and the power analysis was performed with the pwr.chisq.test function of the pwr package. Fisher’s exact test and Cramer’s V were conducted using the Fisher test function and assocstats function, respectively.

Results

Propensity scores were estimated using logistic regression modeling. As shown in Table 9, in comparisons between Groups 1 and 4, Age had a negative effect and Brand Apple had a positive effect; thus, it can be seen that there were many young people and Apple users in Group 4. In addition, the residential area was adjusted. The c-statistic was 0.725, confirming the validity of the model. Models were built in the same way for Groups 1 and 2, and Groups 1 and 3. Matching was performed using estimated propensity scores. Originally, Group 1 had 214 people, Group 2 had 123 people, Group 3 had 109 people, and Group 4 had 116 people. As a result of matching, there were 114 people each in Groups 1 and 2, 91 people each in Groups 1 and 3, and 94 people each in Groups 1 and 4. These sample sizes met the conditions described in Section 3.2. As seen in Table 10, when the respondent attributes of each group after matching are compared, it can be understood that they are homogenized. For example, between Group 1 and Group 4, there was a large difference between Age and Brand Apple, but the values were almost the same. Hypotheses were tested based on data from these matched respondents.

Table 9.

Logistic Regression Model for Estimating Propensity Score.

Variable	Group 1−Group 3				Group 1−Group 4
Variable	Odds ratio	SE	p-Value		Odds ratio	SE	p-Value
(Intercept)	3.745	0.386	.001	***	0.858	0.464	.742
Age	0.563	0.096	.000	***	0.633	0.091	.000***
Area_02_Tohoku					4.470	0.651	.021*
Area_03_Kanto	0.701	0.268	.185		2.081	0.398	.066
Area_04_Chubu					2.479	0.486	.062
Area_05_Kinki					3.005	0.438	.012*
Reason_Design	1.822	0.451	.183
Brand_Apple	1.129	0.285	.671		1.780	0.269	.032*
Brand_Samsung	0.249	0.619	.025	*
Brand_Sony	0.369	0.468	.033	*	0.655	0.414	.306
c-Statistics	0.747				0.725

Note. SE: standard error.

***

p < .001; **p < .01; *p < .05.

Table 10.

Comparison of Respondent Attributes in Both Groups after Propensity Score Matching.

Variable	Group 1		Group 2		Group 1		Group 3		Group 1		Group 4
Variable	Mean	SD	Mean	SD	Mean	SD	Mean	SD	Mean	SD	Mean	SD
Response Device (PC)	0.333	0.473	0.377	0.487	0.440	0.499	0.418	0.496	0.383	0.489	0.468	0.502
Gender (Male)	0.500	0.502	0.465	0.501	0.440	0.499	0.560	0.499	0.479	0.502	0.489	0.503
Age	3.035	1.262	2.974	1.313	2.824	1.322	2.747	1.411	2.968	1.371	2.851	1.495
Area_02_Tohoku	0.026	0.161	0.035	0.185	0.022	0.147	0.044	0.206	0.021	0.145	0.021	0.145
Area_03_Kanto	0.377	0.487	0.404	0.493	0.407	0.494	0.407	0.494	0.436	0.499	0.511	0.503
Area_04_Chubu	0.132	0.340	0.123	0.330	0.176	0.383	0.132	0.340	0.160	0.368	0.128	0.335
Area_05_Kinki	0.219	0.416	0.228	0.421	0.231	0.424	0.209	0.409	0.255	0.438	0.223	0.419
Area_06_Chugoku	0.088	0.284	0.061	0.241	0.044	0.206	0.066	0.250	0.053	0.226	0.064	0.246
Area_07_Shikoku	0.009	0.094	0.026	0.161	0.000	0.000	0.022	0.147	0.000	0.000	0.011	0.103
Area_08_Kyusyu	0.088	0.284	0.044	0.206	0.066	0.250	0.099	0.300	0.043	0.203	0.032	0.177
Income	3.096	1.389	3.263	1.482	3.352	1.417	3.352	1.448	3.362	1.509	3.319	1.497
Purchase_Time	2.447	1.198	2.298	1.030	2.264	1.163	2.418	1.012	2.287	1.188	2.404	1.091
Reason_Design	0.070	0.257	0.070	0.257	0.099	0.300	0.088	0.285	0.064	0.246	0.064	0.246
Reason_Performance	0.219	0.416	0.158	0.366	0.253	0.437	0.231	0.424	0.234	0.426	0.245	0.432
Reason_Price	0.246	0.432	0.289	0.456	0.198	0.401	0.297	0.459	0.266	0.444	0.170	0.378
Reason_Quality	0.079	0.271	0.061	0.241	0.088	0.285	0.044	0.206	0.074	0.264	0.074	0.264
Reason_Reputation	0.009	0.094	0.018	0.132	0.000	0.000	0.011	0.105	0.011	0.103	0.043	0.203
Reason_Usability	0.272	0.447	0.281	0.451	0.275	0.449	0.242	0.431	0.266	0.444	0.245	0.432
Brand_Apple	0.386	0.489	0.377	0.487	0.505	0.503	0.440	0.499	0.500	0.503	0.457	0.501
Brand_Fujitsu	0.026	0.161	0.070	0.257	0.033	0.180	0.066	0.250	0.032	0.177	0.043	0.203
Brand_Samsung	0.096	0.297	0.096	0.297	0.121	0.328	0.044	0.206	0.096	0.296	0.074	0.264
Brand_Sharp	0.158	0.366	0.140	0.349	0.143	0.352	0.198	0.401	0.149	0.358	0.149	0.358
Brand_Sony	0.158	0.366	0.132	0.340	0.066	0.250	0.088	0.285	0.096	0.296	0.128	0.335

SD = standard deviation.

Fisher’s exact test was performed on the matrix of group ×’ distribution for each loyalty index. First, as a reference, the results before propensity score matching were confirmed. As shown in Table 11, there were significant differences between Groups 1 and 3, and Groups 1 and 4. Compared to Group 1, who did not use SNS, it is thought that the fact that there were many young people and a high usage rate for Brand Apple, which has high loyalty, had an effect on each loyalty index. Next, the differences in the response tendencies in each group homogenized by matching were confirmed. As shown in Table 11, there were no significant differences between Groups 1 and 2, which was the same as before matching. Therefore, H1 was supported. Next, there were no significant differences between Groups 1 and 3. This indicated that the significant difference was canceled by homogenizing the conditions, such as respondent attributes and the brand of smartphone owned. Thus, H2 was rejected. However, in Group 1Group 4, the p-values were <.05 for all indices. Since there was still a difference in response tendencies after matching, it could be inferred that this was the effect of Inertia Likes. Hence, H3 was supported. Table 12 shows the matrix of each index in Group 1-Group 4. Looking at the distribution of answers, Group 4 had many answers of “5” for all indices. Comparing the effect sizes of Group 1−Group 3 and Group 1−Group 4, the latter was higher for all indices. In particular, repurchase intention was the highest at 0.272, which indicated a medium effect size. There were no significant differences in Group 1–Group 3. Thus, H4 was supported.

Table 11.

Fisher’s Exact Test Results before and after Matching.

Index	Group1−Group 2			Group1−Group 3			Group1−Group 4
	Before	After		Before	After		Before	After
	p-Value	p-Value	V	p-Value	p-Value	V	p-Value	p-Value	V
Preference	.554	.470	0.126	.028*	.153	0.192	.006**	.045*	0.224
Satisfaction	.746	.686	0.105	.012*	.233	0.178	.002**	.036*	0.227
Repurchase	.419	.776	0.086	.003**	.202	0.182	.000*	.006**	0.272
Recommendation	.178	.650	0.105	.034*	.352	0.157	.000*	.033*	0.232

Note. V = Cramer’s V.

p < .01; *p < .05.

Table 12.

Distribution of Loyalty Indices and Test Results (Group 1−Group 4).

Index	Group	Response					Total	Mean	p-Value
Index	Group	1	2	3	4	5	Total	Mean	p-Value
Preference	Group 1	2	7	40	37	8	94	3.447	.045*
Preference	Group 4	3	3	29	38	21	94	3.755	.045*
Satisfaction	Group 1	1	10	31	46	6	94	3.489	.036*
Satisfaction	Group 4	2	6	18	52	16	94	3.787	.036*
Repurchase	Group 1	3	15	36	31	9	94	3.298	.006**
Repurchase	Group 4	4	5	24	39	22	94	3.745	.006**
Recommendation	Group 1	6	13	49	23	3	94	3.043	.033*
Recommendation	Group 4	4	10	39	26	15	94	3.404	.033*

p < .01; *p < .05.

From the above, hypotheses H1, H3, and H4 were supported, while H2 was rejected. This indicated that even if true Likes become a habit, this would not affect response tendencies for loyalty indices. Furthermore, it was confirmed that, when inertia Likes become a habit, responses tend to have high scores. The starting point of this study was the reduction of the psychological barrier to providing responses indicating favorable attitudes in marketing surveys due to the habituation of inertia. However, although inferior to inertia Likes, it was assumed that the same effect would occur even with the habituation of true Likes; however, there was a clear difference. It was observed that by repeating inertia Likes daily, consumers become “paralyzed” by the act of showing a favorable attitude, leading them to easily provide high scores on marketing surveys.

Discussion

Implications

Being confident enough to recommend a brand to others is a sign of high loyalty. Psychological barriers are higher for the willingness to recommend something to others compared with the satisfaction/preference generated by individual emotions. However, with the spread of SNS in daily life, consumers tend to recommend things to their connected acquaintances using the Like button. Since it is easy to use, many consumers have become accustomed to providing inertia Likes. Therefore, in this study, the focus was on the possibility that Like habits and motives when using SNS reduce the psychological barrier to favorable reactions in marketing surveys. The impact was evaluated using propensity score matching based on the results of an online survey in Japan. As a result, there were no significant differences in the response tendencies of scores for loyalty indices in the group that did not use Likes or the group who used true Likes, compared to the group who did not use SNS. However, significant differences were confirmed in the group who used inertia Likes. Thus, it became clear that the psychological barriers to positive attitudes differed depending on the motivation behind clicking the Like button. It is thought that by repeating inertia Likes on a daily basis, consumers become paralyzed by the act of showing a favorable attitude and thus easily select high scores when taking marketing surveys.

The above results provide important implications for marketing practices. It is common for companies to measure the effects of products/services by regularly observing loyalty indices. However, overlooking the influence of Like habits on survey responses raises concerns that survey results may be biased. In other words, the factor when the survey results fluctuate may not be the influence of products/services, but the influence of changes in the SNS usage tendencies of the respondent group, especially regarding Like motivation. In this case, the company may make a mistake in decision–making. Similarly, in academic research, biases such as SNS habits and motives should be considered, so that effects other than those intended are not included. In addition, the results encourage marketers to consider Like habits when using brand attitude metrics in identifying high–loyalty customers. Otherwise, there is a risk of overestimating customers who are just paralyzed by Like’s habits. In this way, it is significant to grasp the usage status of SNS in terms of both measuring the effectiveness of marketing measures and identifying loyal customers. This study provides useful suggestions for both industrial and academic research.

Limitations and Future Work

This study also has several limitations. First, since this study targeted only the Japanese market, there is a limit to the generalization of the findings. In the future, the scope should be expanded to other countries, and differences in Like motives and the related impact on survey responses should be examined. Second, the effect of Like frequency was not considered. As shown in Tables 5 and 6, the sample size of each group was approximately 100, making it difficult to subdivide by frequency. Third, the influence of Post was not considered. As shown in Table 3, the active ratios were Read (64.3%) > Like (43.8%) > Post (35.2%). There are very few people who Post but do not Like, and those people have not been evaluated. However, in the real world, such people are considered a minority. Fourth, Read/Like/Post use data was based on awareness surveys rather than behavioral data. Ideally, research should track usage records and handle factual data. However, this method is difficult to implement due to privacy considerations. Fifth, the effect of changing Likes to another similar function is unclear. It has been pointed out that adherence to Like has become too strong and has a negative impact on consumers’ mental health (Royal Public Society for Public Health, 2019). A quantified metric as a Like count stimulates the desire for more support and intensifies competition with peers. There is concern that this competitive thinking will increase psychological burden, especially among young people (Li et al., 2018). In response to these social issues, Instagram began testing by hiding Like counts in 2019 (Wagner, 2019), while Facebook has announced that it will abolish Like and focus on Follow (Awan, 2021). However, there are many dissenting opinions about this choice, and a browser plugin that restores the hidden Like count has been unofficially developed (Leighton, 2019). As of January 2021, in Japan, the Like function has been implemented on Twitter, and TikTok, and the Like button still exists on Instagram (the Like count is hidden). However, although there is a possibility that the name will be changed from Like or the function will be modified in the future, it is unlikely that it will be eliminated. If there is a change in functionality, this verification should be performed again. Sixth, the subject of Like is not considered in this study; there are various targets and reasons for pressing the Like button, such as friends, celebrities, and brands. Differences in the effects of these differences on paralysis have not been clarified yet. These may serve as future research topics.

The impact of Likes on consumer behavior due to the rapid penetration of SNS has attracted attention. However, Like motives are often unconscious or are related to the maintenance of social relationships. In this study, it was found that the ratio of true Likes and inertia Likes was almost the same. Therefore, a critical examination of Likes without favorable attitudes has come to be discussed (Burke & Kraut, 2014; John et al., 2017). To the best of my knowledge, this study was the first to focus on the effects of inertia Likes on response trends in marketing research. Research on factors affecting survey quality have been extensively conducted, such as questionnaire length (Galesic & Bosnjak, 2009), screen design (Healey, 2007; Maloshonok & Terentev, 2016), incentives for respondents (Kato et al., 2020), and answering devices (Lugtig & Toepoel, 2016; Mavletova, 2013). However, the perspective presented by this study was missing in the current literature. Based on the verification results, marketers were urged to understand Like’s habits on social media, both in measuring the effectiveness of marketing measures and in identifying the loyal customers. Thus, this study contributes to the development of research on the effectiveness of Likes on SNS and the quality of marketing research.

Footnotes

Declaration of Conflicting Interests

The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Takumi Kato

References

Aaker

D. A.

(1991). Managing brand equity. The Free Press.

Aaker

D. A.

Joachimsthaler

(2000). Brand leadership. Free Press.

Assael

(1987). Consumer behavior and marketing action. Kent Publishing.

Awan

(2021, January 6). Introducing the new pages experience. Facebook. https://about.fb.com/news/2021/01/introducing-the-new-page-experience

Bastos

M. T.

(2015). Shares, pins, and tweets: News readership from daily papers to social media. Journalism Studies, 16(3), 305–325. https://doi.org/10.1080/1461670x.2014.891857

Beukeboom

C. J.

Kerkhof

de Vries

(2015). Does a virtual like cause actual liking? How following a brand’s Facebook updates enhances brand evaluations and purchase intention. Journal of Interactive Marketing, 32, 26–36. https://doi.org/10.1016/j.intmar.2015.09.003

Burke

Kraut

R. E.

(2014). Growing closer on Facebook: Changes in tie strength through social network site use [Conference session]. Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, pp. 4187–4196. Association for Computing Machinery (ACM). https://doi.org/10.1145/2556288.2557094.

Chin

C. Y.

H. P.

C. M.

(2015). Facebook users’ motivation for clicking the “like” button. Social Behavior and Personality: an international journal, 43(4), 579–592. https://doi.org/10.2224/sbp.2015.43.4.579

Cohen

(1992). A power primer. Psychological Bulletin, 112(1), 155–159. https://doi.org/10.1037/0033-2909.112.1.155

10.

Cushing

Douglas-Tate

(1985). The effect of people/product relationships on advertising processing. In Alwitt

L. F.

Mitchell

A. A.

(Eds.), Psychological processes and advertising effects (pp. 241–259). Lawrence Erlbaum Associates.

11.

Dick

A. S.

Basu

(1994). Customer loyalty: Toward an integrated conceptual framework. Journal of the Academy of Marketing Science, 22(2), 99–113. https://doi.org/10.1177/0092070394222001

12.

Ding

Cheng

H. K.

Duan

Jin

(2017). The power of the “like” button: The impact of social media on box office. Decision Support Systems, 94, 77–84. https://doi.org/10.1016/j.dss.2016.11.002

13.

Egebark

Ekström

(2018). Liking what others “like”: Using Facebook to identify determinants of conformity. Experimental Economics, 21(4), 793–814. https://doi.org/10.1007/s10683-017-9552-1

14.

Eranti

Lonkila

(2015). The social significance of the Facebook Like button. First Monday, 20(6), 1–14. https://doi.org/10.5210/fm.v20i6.5505

15.

Facebook. (n.d). Reactions. https://en.facebookbrand.com/facebookapp/assets/reactions/

16.

Galesic

Bosnjak

(2009). Effects of questionnaire length on participation and indicators of response quality in a web survey. Public Opinion Quarterly, 73(2), 349–360. https://doi.org/10.1093/poq/nfp031

17.

Gavilanes

J. M.

Flatten

T. C.

Brettel

(2018). Content strategies for digital consumer engagement in social networks: Why advertising is an antecedent of engagement. Journal of Advertising, 47(1), 4–23. https://doi.org/10.1080/00913367.2017.1405751

18.

Gelles

(2010, September 21). E-commerce takes instant liking to Facebook button. Financial Times. https://www.ft.com/content/1599be2e-c5a9-11df-ab48-00144feab49a

19.

Gerlitz

Helmond

(2013). The like economy: Social buttons and the data-intensive web. New Media & Society, 15(8), 1348–1365. https://doi.org/10.1177/1461444812472322

20.

Gerodimos

Justinussen

(2015). Obama’s 2012 Facebook campaign: Political communication in the age of the like button. Journal of Information Technology & Politics, 12(2), 113–132. https://doi.org/10.1080/19331681.2014.982266

21.

Gilliland

D. I.

Bello

D. C.

(2002). Two sides to attitudinal commitment: The effect of calculative and loyalty commitment on enforcement mechanisms in distribution channels. Journal of the Academy of Marketing Science, 30(1), 24–43. https://doi.org/10.1177/03079450094306

22.

Gonçalves

F. M. R. R.

Cândido

C. J. F.

Feliciano

I. M. P. L.

(2020). Inertia, group conformity and customer loyalty in healthcare in the information age. Journal of Service Theory and Practice, 30(3), 307–330. https://doi.org/10.1108/jstp-08-2019-0184

23.

Grant

J. E.

Chamberlain

S. R.

Schreiber

L. R. N.

Odlaug

B. L.

(2012). Gender-related clinical and neurocognitive differences in individuals seeking treatment for pathological gambling. Journal of Psychiatric Research, 46(9), 1206–1211. https://doi.org/10.1016/j.jpsychires.2012.05.013

24.

Hayes

R. A.

Carr

C. T.

Wohn

D. Y.

(2016). One click, many meanings: Interpreting paralinguistic digital affordances in social media. Journal of Broadcasting & Electronic Media, 60(1), 171–187. https://doi.org/10.1080/08838151.2015.1127248

25.

Healey

(2007). Drop downs and scroll mice: The effect of response option format and input mechanism employed on data quality in web surveys. Social Science Computer Review, 25(1), 111–128. https://doi.org/10.1177/0894439306293888

26.

Hortaçsu

Madanizadeh

S. A.

Puller

S. L.

(2017). Power to choose? An analysis of consumer inertia in the residential electricity market. American Economic Journal Economic Policy, 9(4), 192–226. https://doi.org/10.1257/pol.20150235

27.

Y. G.

North

Liu

(2017). Staking reputation on stakeholders: How does stakeholders’ Facebook engagement help or ruin a company’s reputation? Public Relations Review, 43(1), 201–210. https://doi.org/10.1016/j.pubrev.2016.12.004

28.

John

L. K.

Emrich

Gupta

Norton

M. I.

(2017). Does “liking” lead to loving? The impact of joining a brand’s social network on marketing outcomes. JMR, Journal of Marketing Research, 54(1), 144–155. https://doi.org/10.1509/jmr.14.0237

29.

Kabadayi

Price

(2014). Consumer – Brand engagement on Facebook: Liking and commenting behaviors. Journal of Research in Interactive Marketing, 8(3), 203–223. https://doi.org/10.1108/jrim-12-2013-0081

30.

Kato

Kishida

Umeyama

Jin

Tsuda

(2020). A random extraction method with high market representation for online surveys. International Journal of Business Innovation and Research, 22(4), 569–584. https://doi.org/10.1504/ijbir.2020.109036

31.

Kemp

(2022, January 26). Digital 2022: Global over view report. Datareportal. https://datareportal.com/reports/digital-2022-global-overview-report

32.

Khalilzadeh

Tasci

A. D. A.

(2017). Large sample size, significance level, and the effect size: Solutions to perils of using big data for academic research. Tourism Management, 62, 89–96. https://doi.org/10.1016/j.tourman.2017.03.026

33.

Khrais

(2018, April 11). The most powerful tool in social media. Marketplace. https://www.marketplace.org/2018/04/11/it-was-known-button-it-was-awesome-button/

34.

Kim

Kang

(2016). How user loyalty and nonconscious inertia influence the continued use of mobile communications platforms. International Journal of Mobile Communications, 14(4), 387–410. https://doi.org/10.1504/ijmc.2016.077337

35.

Kim

Yang

S. U.

(2017). Like, comment, and share on Facebook: How each behavior differs from the other. Public Relations Review, 43(2), 441–449. https://doi.org/10.1016/j.pubrev.2017.02.006

36.

Kim

J. W.

(2014). Scan and click: The uses and gratifications of social recommendation systems. Computers in Human Behavior, 33, 184–191. https://doi.org/10.1016/j.chb.2014.01.028

37.

Kono

(2009). From the marketers’ perspective: The interactive media situation in Japan. In Gerbarg

(Ed.), Television goes digital (pp. 57–59). Springer.

38.

Kosinski

Stillwell

Graepel

(2013). Private traits and attributes are predictable from digital records of human behavior. Proceedings of the National Academy of Sciences of the United States of America, 110(15), 5802–5805. https://doi.org/10.1073/pnas.1218772110

39.

Krosnick

J. A.

(1991). Response strategies for coping with the cognitive demands of attitude measures in surveys. Applied Cognitive Psychology, 5(3), 213–236. https://doi.org/10.1002/acp.2350050305

40.

Kudeshia

Sikdar

Mittal

(2016). Spreading love through fan page liking: A perspective on small scale entrepreneurs. Computers in Human Behavior, 54, 257–270. https://doi.org/10.1016/j.chb.2015.08.003

41.

Kuo

Yang

(2013). Effects of inertia and satisfaction in female online shoppers on repeat-purchase intention: The moderating roles of word-of-mouth and alternative attraction. Managing Service Quality: An International Journal, 23(3), 168–187. https://doi.org/10.1108/09604521311312219

42.

Lee

S. Y.

Hansen

S. S.

Lee

J. K.

(2016). What makes us click “like” on Facebook? Examining psychological, technological, and motivational factors on virtual endorsement. Computer Communications, 73, 332–341. https://doi.org/10.1016/j.comcom.2015.08.002

43.

Leighton

(2019, December 10). This chrome extension brings Instagram posts’ likes back. Forbes. https://www.forbes.com/sites/heatherleighton/2019/12/10/this-chrome-extension-brings-instagram-posts-likes-back/?sh=241f52fe146c

44.

Leiner

D. J.

Kobilke

Rueß

Brosius

H. B.

(2018). Functional domains of social media platforms: Structuring the uses of Facebook to better understand its gratifications. Computers in Human Behavior, 83, 194–203. https://doi.org/10.1016/j.chb.2018.01.042

45.

Levordashka

Utz

Ambros

(2021). What’s in a like? Motivations for pressing the like button. International AAAI Conference on Web and Social Media, 10(1), 623–626.

46.

Chang

Chua

T. H. H.

Loh

R. S. M.

(2018). Likes” as KPI: An examination of teenage girls’ perspective on peer feedback on Instagram and its influence on coping response. Telematics and Informatics, 35(7), 1994–2005. https://doi.org/10.1016/j.tele.2018.07.003

47.

Lipsman

Mudd

Rich

Bruich

(2012). The power of “like”: How brands reach (and influence) fans through social-media marketing. Journal of Advertising Research, 52(1), 40–52. https://doi.org/10.2501/jar-52-1-040-052

48.

Liu

Y. G.

North

Yang

(2017). Like it or not: The Fortune 500’s Facebook strategies to generate users’ electronic word-of-mouth. Computers in Human Behavior, 73, 605–613. https://doi.org/10.1016/j.chb.2017.03.068

49.

López

Sicilia

Verlegh

P. W. J.

(2017). Click like if you like it”: The effect of directional posts on social network sites. Online Information Review, 41(5), 672–690. https://doi.org/10.1108/oir-01-2016-0020

50.

Lugtig

Toepoel

(2016). The use of PCs, smartphones, and tablets in a probability-based panel survey: Effects on survey measurement error. Social Science Computer Review, 34(1), 78–94. https://doi.org/10.1177/0894439315574248

51.

Maloshonok

Terentev

(2016). The impact of visual design and response formats on data quality in a web survey of MOOC students. Computers in Human Behavior, 62, 506–515. https://doi.org/10.1016/j.chb.2016.04.025

52.

Mavletova

(2013). Data quality in PC and mobile web surveys. Social Science Computer Review, 31(6), 725–743. https://doi.org/10.1177/0894439313485201

53.

MM Research. (2021, February 10). SIM fleece smartphone surges with the start of Rakuten MNO service. https://www.m2ri.jp/release/detail.html?id=473 [in Japanese].

54.

Moffat

(2019). The power of likes on social media: Friend or foe? Future of Customer Engagement and Experience. https://www.the-future-of-commerce.com/2019/10/07/the-power-of-likes-on-social-media/

55.

Munar

A. M.

Jacobsen

J. K. S.

(2014). Motivations for sharing tourism experiences through social media. Tourism Management, 43, 46–54. https://doi.org/10.1016/j.tourman.2014.01.012

56.

Naylor

R. W.

Lamberton

C. P.

West

P. M.

(2012). Beyond the “like” button: The impact of mere virtual presence on brand evaluations and purchase intentions in social media settings. Journal of Marketing, 76(6), 105–120. https://doi.org/10.1509/jm.11.0105

57.

Oliver

R. L.

(1999). Whence consumer loyalty? Journal of Marketing, 63(4), 33–44. https://doi.org/10.2307/1252099

58.

Pöyry

Parvinen

Malmivaara

(2013). Can we get from liking to buying? Behavioral differences in hedonic and utilitarian Facebook usage. Electronic Commerce Research and Applications, 12(4), 224–235. https://doi.org/10.1016/j.elerap.2013.01.003

59.

Rosenbaum

P. R.

Rubin

D. B.

(1983). The central role of the propensity score in observational studies for causal effects. Biometrika, 70(1), 41–55. https://doi.org/10.1093/biomet/70.1.41

60.

Royal Public Society for Public Health. (2019, August 30). Go “Scroll Free” this September and save 100 hours of your life. https://www.rsph.org.uk/about-us/news/go-scroll-free-this-september-and-save-100-hours-of-your-life.html

61.

Schumacher

Kent

(2020), April 2). 8 charts on internet use around the world as countries grapple with COVID-19. Pew Research Center. https://www.pewresearch.org/fact-tank/2020/04/02/8-charts-on-internet-use-around-the-world-as-countries-grapple-with-covid-19/

62.

Sherman

L. E.

Payton

A. A.

Hernandez

L. M.

Greenfield

P. M.

Dapretto

(2016). The power of the like in adolescence: Effects of peer influence on neural and behavioral responses to social media. Psychological Science, 27(7), 1027–1035. https://doi.org/10.1177/0956797616645673

63.

Sorensen

Andrews

Drennan

(2017). Using social media posts as resources for engaging in value co-creation: The case for social media-based cause brand communities. Journal of Service Theory and Practice, 27(4), 898–922. https://doi.org/10.1108/jstp-04-2016-0080

64.

Sudarshan. (2021, February 25). Apple was Japan’s largest smartphone brand in 2020 with a 47.3% market share. Gizmochina. https://www.gizmochina.com/2021/02/25/japan-mobile-phone-smartphone-market-q4-2020-idc/#:∼:text=As%20for%20the%20year%202020,and%207.5%25%20market%20shares%20respectively

65.

Sumner

E. M.

Ruge-Jones

Alcorn

(2018). A functional approach to the Facebook lLike button: An exploration of meaning, interpersonal functionality, and potential alternative response buttons. New Media & Society, 20(4), 1451–1469. https://doi.org/10.1177/1461444817697917

66.

Swani

Milne

G. P.

Brown

(2013). Spreading the word through likes on Facebook. Journal of Research in Interactive Marketing, 7(4), 269–294. https://doi.org/10.1108/jrim-05-2013-0026

67.

Thai

T. D. H.

Wang

(2020). Investigating the effect of social endorsement on customer brand relationships by using statistical analysis and fuzzy set qualitative comparative analysis (fsQCA). Computers in Human Behavior, 113, 106499. https://doi.org/10.1016/j.chb.2020.106499

68.

Thompson

Wang

Daya

(2020). Determinants of news sharing behavior on social media. Journal of Computer Information Systems, 60(6), 593–601. https://doi.org/10.1080/08874417.2019.1566803

69.

Torgerson

(2008). Designing randomised trials in health, education and the social sciences: An introduction. Palgrave MacMillan.

70.

Trattner

Kappe

(2013). Social stream marketing on Facebook: A case study. International Journal of Social and Humanistic Computing, 2(1/2), 86–103. https://doi.org/10.1504/ijshc.2013.053268

71.

VanMeter

Syrdal

H. A.

Powell-Mantel

Grisaffe

D. B.

Nesson

E. T.

(2018). Don’t just “like” me, promote me: How attachment and attitude influence brand related behaviors on social media. Journal of Interactive Marketing, 43, 83–97. https://doi.org/10.1016/j.intmar.2018.03.003

72.

Wagner

(2019), November 9). Instagram will remove ‘likes’ from posts for some U.S. Users. Bloomberg. https://www.bloomberg.com/news/articles/2019-11-09/instagram-will-remove-likes-from-posts-for-some-u-s-users

73.

Wang

Wei

(2012). Social media peer communication and impacts on purchase intentions: A consumer socialization framework. Journal of Interactive Marketing, 26(4), 198–208. https://doi.org/10.1016/j.intmar.2011.11.004

74.

Wang

X. W.

Cao

Y. M.

Park

(2019). The relationships among community experience, community commitment, brand attitude, and purchase intention in social media. International Journal of Information Management, 49, 475–488. https://doi.org/10.1016/j.ijinfomgt.2019.07.018

75.

White

Yanamandram

(2004). Why customers stay: Reasons and consequences of inertia in financial services. Managing Service Quality: An International Journal, 14(2/3), 183–194. https://doi.org/10.1108/09604520410528608

76.

Yoon

North

Hong

Liu

(2018). Attracting comments: Digital engagement metrics on Facebook and financial performance. Journal of Advertising, 47(1), 24–37. https://doi.org/10.1080/00913367.2017.1405753

77.

Youyou

Kosinski

Stillwell

(2015). Computer-based personality judgments are more accurate than those made by humans. Proceedings of the National Academy of Sciences of the United States of America, 112(4), 1036–1040. https://doi.org/10.1073/pnas.1418680112