Descriptive social norms and resource cues influence choice by additive and separate effects

Abstract

Descriptive social norms have attracted much attention in social influence research. Regarding consumer choice, it is however unclear if, and to what extent, the influence of social norms is related to resource-state information. In two experiments, including 384 and 724 participants, respectively, we assess the unique and combined effects of these influences on both choice and preferences. Results showed consistent effects of descriptive social norms, influencing both choice and preferences across the two experiments. When a resource cue was provided in Experiment 1, a small non-significant difference compared to the control condition indicated that information about resource states might affect choice. This effect was replicated with statistical significance in Experiment 2. No effect of such a cue was detected on preferences in either experiment. Present results suggest that the effects of descriptive social norms and resource cues are independent and additive.

Keywords:

• social influence
• social norms
• resource cues
• decision-making

1. Introduction

Research on the dynamics behind social influence has been applied to promote various consumer behaviors, both within health related (Mattern & Neighbors, 2004; Robinson et al., 2014), and pro-environmental domains (Abrahamse & Steg, 2013; Allcott, 2011; Schultz et al., 2007). However, since the psychological effects of social influence vary substantially (e.g. Abrahamse & Steg, 2013) it is important to investigate how different sources of social influence affect consumer behavior. One way to assess this is to examine combinations of social influence approaches. Previous research has, for example, reported that coupling a social norms-based intervention with commitments promoted water conservation (Jaeger & Schultz, 2017), while others have found that coupling social norms with a contest-based intervention decreased the long-term effects of an energy conservation intervention (Alberts et al., 2016).

In the present research, we test the novel combination of descriptive social norms (signaling other people’s behavior), and cues about resource states (signaling possible resource scarcity). We explore if resource cues, here operationally defined as “shelf-based scarcity” that signals relative stocking level depletion, can add to the influential power of descriptive social norms. We choose to focus on such cues for two reasons. First, resource cues are an ecologically valid social influence approach (i.e., can be found or implemented in real-world purchase contexts) that has attracted much research and shown a general effect on both purchase intentions and behavior (see Oruc, 2015). Second, resource cues are similar to descriptive social norms, in how both implicitly inform the consumer about other people’s behaviors. Across two experiments, we will test whether cues of resource states can add to the influence of descriptive social norms, by assessing choices and preferences in online choice tasks.

1.1. The influence of descriptive social norms

Social norms can be defined as “rules or standards that are understood by members of a group, and that guide and/or constrain social behavior without the force of law” (Cialdini & Trost, 1998, p.152). Norms can be further divided into personal norms, injunctive social norms, and descriptive social norms. These norms differ in motivational basis as personal norms are driven by people’s intrinsic obligation, injunctive norms by other people’s (dis)approval, and descriptive norms guide behavior by observing other people’s behaviors and signs thereof (see Thøgersen, 2006). Descriptive social norms have been shown to promote various behaviors (Bergquist et al., 2019; Burger & Shelton, 2011; Scheibehenne et al., 2016). One possible explanation behind the influential power of descriptive norms is the goal of accuracy (Cialdini & Goldstein, 2004). This refers to how people may use others’ behaviors as guidelines for assessing “correct” or “adaptive” behaviors (“goals”) in a specific situation.

Interventions aiming to promote behavioral change using social norms-based information have been shown to successively promote both pro-health (Robinson et al., 2014; Sheeran et al., 1999) and pro-environmental behaviors (Bergquist et al., 2019). Although social norms-based interventions have shown to affect behaviors in general, some studies report no effect (e.g., Richter et al., 2018; Scheibehenne et al., 2016). To the best of our knowledge, no past research has assessed the relationship between social norms and resource cues.

1.2. The influence of resource cues

In contrast to an objective state of a resource, the scarcity concept assesses a perceived resource state. In general terms, scarcity can then be defined as “having less than you feel you need” (Mullainathan & Shafir, 2013, p.4). In order to sustain vital resources, it is necessary to assess whether resources are scarce in relation to specific needs (Hobfoll, 1989, 2001, 2002). Individuals then rely on resource assessments to mobilize adequate coping strategies (Folkman et al., 1986; Lazarus & Folkman, 1984). In line with previous research showing more general effects such as loss aversion (Thaler et al., 1997; Tom et al., 2007, Tversky & Kahneman, 1991), individuals who receive information that signals resource scarcity are expected to become increasingly defensive of their resources. Other research furthermore shows that people use relative information as cues to make resource assessments, and that this can explain their subjective well-being (Einarsdóttir et al., 2018b). People thus depend on points of reference when evaluating whether a resource is scarce. Also in daily life, when choosing consumer products, do people anchor their evaluations on resource cues available in the immediate decision context (Gigerenzer, Todd & The ABC group, 1999).

We find this reasoning particularly applicable in contexts where resource scarcity can be seen as product scarcity, defined as “a condition or message that communicates a certain or potential unavailability of a product in the future along with the availability of a product in the present, all of which are directed at all possible recipients of a product.” (Oruc, 2015, p 3). According to this definition, scarcity is a specific type of unavailability, where a product is available at present but certainly or possibly unavailable in the future. Moreover, Gierl et al. (2008) distinguish scarcity due to supply (e.g., limited edition, seasonal restriction) from scarcity due to demands (e.g., shelf-based scarcity). Importantly, only the latter implies consumer behaviors, as it implicitly signals that others have chosen that product over an alternative product. This distinction is important for the present research, as it suggests that both shelf-based scarcity and descriptive social norms can exert social influence based on direct or indirect cues of other people’s behaviors. Research on shelf-based scarcity shows that people see a scarce product as more popular and of higher quality compared to an abundant product (van Herpen et al., 2014). Hence, both shelf-based scarcity and descriptive social norms are signaling “popularity”, which makes the comparison and possible combination of these social influence techniques especially important.

1.3. The present research

The SARS-COV-2 pandemic has changed consumption patterns. In many countries, people have for instance hoarded toilet paper, which has been ridiculed and described as irrational and puzzling. In this research, we question if people choose certain products because they feel that the supply is running out. Is the sight of empty shelves seen as a signal of a vital resource being threatened? Or do people choose these products because they see other people do so? These questions lead us to investigate the unique and combined effects of descriptive social norms and resource scarcity cues.

In two experiments where participants chose between two types of toilet papers, we manipulated resource cues by using the shelf-based scarcity research paradigm (e.g., Park & Lehmann, 2011), signaling relative stocking level depletion by changing the number of products available, showing either equal or unequal piles of products. Descriptive social norms were manipulated by assigning participants to a group (without knowing that other group members were programmed avatars) in which the majority of the group members chose uniformly.

1.3.1. Hypotheses

We hypothesized two main effects, one for the descriptive social norm and one for the resource cue. Firstly, we hypothesized a main effect of the descriptive social norm (H1). Secondly, given the research presented above, we expected the resource cue to affect participants’ purchases. We tested two competing hypotheses. In line with the meta-analytic overall effect, showing that resource scarcity promoted purchase intention, r = .19, 95% CI [.14, .24] (Oruc, 2015), we hypothesized that the resource cue could make participants choose more from the smaller pile (here the extra-long toilet paper option) (H2a). Due the meta-analytic dispersion, showing that 26.5% of the primary studies actually report the reversed effect, and that the 95% CI included the negative effect in 61% of primary studies (Oruc, 2015), we also provide the contrasting hypothesis that the resource cue could make participants choose more from the larger pile (here the extra-soft toilet paper option) (H2b).

We will further explore the interrelationship between descriptive social norms and resource cues. Specifically, if both these manipulations caused positively directed effects (in the same direction) we would expect more participants in the resource cue and norm condition (group D) to choose the extra-long option, compared to both the resource cue condition and the norm condition. If the effects of descriptive social norms and resource cues prove to be independent as well as positive, we expect to see an increased effect in group D where these manipulations are combined. However, if descriptive social norms and resource cues show effects in opposite directions, we would expect these effects to be canceled out, at least partly. Furthermore, if the effect of one of these factors depends on the effect of the other factor, we would expect to see a significant descriptive social norm by resource cue interaction.

To further examine our research question, we measure how preferences are affected in all conditions. As resource cues have been described as triggers of automatic processing (Ariely & Wertenbroch, 2002; Gigerenzer, Todd & The ABC group, 1999; Shah et al., 2012), the effects of descriptive social norms are driven by the goal of accuracy (Cialdini & Goldstein, 2004). Additionally, Oruc’s meta-analysis on scarcity (2015) found no effect of attitude towards the product. We thus hypothesize (H3) stronger positive preferences for participants in the descriptive social norm condition compared to the resource cue and control conditions.

2. Experiment 1

2.1. Methods

2.1.1. Participants

A power calculation based on the expected effect of descriptive norms (d = 0.32: Bergquist et al., 2019), suggested a sample of 430 participants (F = 0.16, α = .05, β = .80, df = 3). We recruited 443 participants located in the USA using Amazon’s Mechanical Turk (MTurk). After excluding inattentive participants (participants failing to report what product they chose) the final sample was N = 384 (64.1% male).

2.1.2. Design

To assess the influence of descriptive social norms and resource cues on product choice, we set up a 2 (descriptive norm: no vs. yes) × 2 (resource cue: no vs. yes) between-groups design, randomly assigning participants to either a control condition (group A), a resource cue condition (group B), a descriptive social-norm condition (group C), or a resource cue and norm condition (group D).

Participants were informed that the study was “a choice task”, that “there are no right or wrong answers”, and to “imagine that you are in your local grocery store and intend to buy two products: bottled water and toilet paper”. Next, to conceal the purpose of the study participants performed a practice task before the main task, also including a pair of grocery products.

2.1.3. Measures

In the first task, all participants were presented with two soft drinks, one yellow and one green and asked to choose one of these. No information about descriptive social norms or resource states was provided in this task.

Participants were then presented with the main task in which they were asked to choose between two types of toilet paper. We used the same type of product adding labels to provide a rationale for choosing between the products: “extra soft” or “extra-long”. The descriptive social norm conditions were designed to promote the extra-long option. The conditions involving resource cues (group B and D) were designed such that the displayed number of extra-long paper rolls were few (in the control condition, 38.1% choose “extra soft”).

In both the practice task and the main task, participants were presented with the two alternatives and asked “Which bottled water/toilet paper do you buy?” Next, participants stated their preference by answering the question “How strongly do you prefer your chosen product?” followed by a scale ranging from 1 “I strongly prefer the green soft drink/extra soft toilet paper”, to 7 “I strongly prefer the yellow soft drink/extra-long toilet paper”, also including a neutral preference 4 “I have no preference”.

2.1.4. Procedure

In Experiment 1, the resource cue was operationalized in accordance with the shelf-based scarcity research paradigm, by manipulating the resource state making the “extra-long” paper scarce (hence the “extra-long” paper came in a smaller pile and the “extra soft” paper in a larger pile). More specifically, when participants in the resource cue condition made their choice, the stimulus material displayed 2 packages of “extra-long” papers and 7 packages of “extra soft” papers (group B).

The descriptive social norm was manipulated in three steps: (1) participants were presented with seven avatars and asked to select and name a character for the main choice task; (2) participants were informed that they had been assigned to a group of six other participants, revealing others participants avatars and names, and finally; (3) participants were provided with descriptive social normative information by showing other avatars’ choices of the extra-long option, one at a time. In this descriptive social norm condition, the resource state was held constant throughout the procedure (group C).

Participants assigned to the combined resource cue and norm condition went through the same experimental procedure as in the descriptive social norm condition. But they also received information about the resource state by one item of “extra-long” toilet paper being visibly removed as each avatar made his or her choice (group D). The decision context thus entailed scarcity of the extra-long option as well as a descriptive norm of all other group members choosing the same option.

Finally, participants assigned to the control condition received neither the resource cue nor the descriptive norm (group A).

After finalizing the main choice task, all participants were asked to report which product they chose to ensure that they paid attention, asked to report their gender, and finally debriefed about the experimental design.

2.2. Results

2.2.1. Choice

Results of a logistic regression with choice as the dependent variable shows support for the main effect of descriptive norm (H1). Model 1.2, which includes the resource cue by descriptive norm interaction term, shows a non-significant effect of resource (See Table 1). In descriptive terms, results showed a somewhat weak conditional effect indicating that the larger pile (here “extra soft”) might have promoted choice. These results are descriptively in line with H2b rather than H2a. No interaction effects were detected (see Table 1 for details).

Table 1 Descriptive statistics for choice (% choosing “extra-long”) in Experiment 1.

	Experiment 1
	Model 1.1			Model 1.2
Independent variables	B	SE	p	B	SE	p
Constant	−0.35	0.18	0.054	−0.24	0.21	0.255
Descriptive norm	−0.52	0.21	0.014	−0.30	0.29	0.309
Scarcity cue	−0.27	0.21	0.197	0.52	0.31	0.093
Descriptive norm × Scarcity cue				−0.47	0.43	0.270
Nagelkerke R^2	0.027			0.031
−2 Log-likelihood	504.2			503.0
N	384

Notes. Bs are unstandardized regression coefficients (logged-odds effects) with associated standard errors. Please note that the descriptive social norm manipulation was aligned with a scarcity resource cue.

2.2.2. Preference

To further test the effects on product preferences, we ran a factorial 2 × 2 ANOVA with descriptive social norm and resource cue as independent variables on rating of preference (Table 2). In line with H3, results revealed a main effect of norms (F(1, 380) = 9.77, p = .01), showing that preferences were stronger in the norm condition (M = 4.23, SD = 2.08) compared to the control condition (M = 3.63, SD = 2.15, d = 0.29). No main effect of resource cue (F(1, 380) = 0.01, p = .92) and no interaction effect (F(1, 380) = 0.06, p = .80) was found.

Table 2 Descriptive statistics for preference in Experiment 1 including sample sizes, means and standard deviations, and effect sizes d comparing each condition to the control condition.

Condition	Sample size	Mean (sd)	Effect size d (95% CI)
Control	97	3.63 (2.15)
Descriptive norm	97	4.23 (2.08)	0.29 (0.06, 0.57)
Scarcity cue	97	3.60 (1.95)	−0.01 (-0.29, 0.27)
Descriptive norm and Scarcity cue	93	4.30 (1.98)	0.33 (0.04, 0.62)

2.3. Discussion

Experiment 1 found the predicted effect of descriptive social norm, and a non-significant effect for resource cue. In descriptive terms, the latter non-significant effect indicated that participants might have been influenced by the larger pile, although this should be interpreted with caution. One explanation for this non-significant effect is that Experiment 1 ended up being somewhat under-powered. Therefore, we will try to replicate these effects in Experiment 2, now with increased statistical power.

We found no interaction effect between the descriptive social norm and resource cue. This might be explained by the manipulations promoting different choice alternatives (i.e., a descriptive social norm promoted the “extra-long”, while the resource cue promoted the “extra soft”). To once again test the relationships between these effects, we will conduct Experiment 2, now aligning the choice effects of the descriptive social norm and resource cue.

In line with Hypothesis 3, we found that preferences were affected in the descriptive social norms condition, but not in the resource cue condition. These results are in line with past research (Oruc, 2015). We aim to replicate this effect in Experiment 2. One explanation for this effect is that the influence of descriptive social norms is driven by the goal of accuracy (Cialdini & Goldstein, 2004). It is possible that participants interpreted the normative information as “social proof” suggesting that other people’s choices were interpreted as the “right” or “adaptive” behavior. Consequently, participants adjusted their preferences for the approved alternative. By contrast, the influence of resource cue might be more of an automatic reaction that bypasses preferences.

3. Experiment 2

In line with H2b, Experiment 1 indicated that when presented with a resource cue, more participants chose from the larger rather than the smaller pile of toilet paper. Experiment 2 was designed to replicate this effect, this time by comparing congruent choice alternatives: the larger resource option was now aligned with the descriptively normative option. Hence, the design was once again a 2 (descriptive norm: no vs. yes) × 2 (resource cue: no vs. yes) between groups design, randomly assigning participants to either a control condition (group A), a resource cue condition (group B), a descriptive social norm condition (group C), or a resource cue and norm condition (group D). Because Experiment 1 was limited by ending up with somewhat fewer participants than planned, we increased the number of participants to boost statistical power.

To further investigate the assumed effects of resource cue, individual differences in terms of childhood resource states and current resource states were measured. Previous research suggests that resource scarcity during childhood can affect resource relevant behaviors later in life. After experiencing scarcity, people who grew up poor tend to increase immediate gain-seeking behaviors, whereas those who grew up rich instead tend to reduce these behaviors (Griskevicius et al., 2013; Hansla & Johansson, 2020; Kenrick et al., 2009; Mittal & Griskevicius, 2016). According to the so-called sensitization hypothesis (Ellis et al., 2017), such behaviors can be manifested in adulthood when activated by cues of resource scarcity. A measure of childhood scarcity was added in order to explore such effects.

3.1. Hypotheses

Again, we hypothesized two main effects, one for the descriptive social norm and one for the resource cue. As a first hypothesis (H1), we predicted a main effect of descriptive social norm showing that after being exposed to the descriptive social norm promoting this choice, more participants would choose the extra-long option compared to the control condition. Given the results from Experiment 1, we here retained hypothesis H2b predicting that the resource cue would affect the larger pile of toilet paper. We will also test if preferences are affected only by the descriptive social norm and not by the resource cue (H3). In addition, the effects of childhood scarcity will be explored.

3.2. Methods

3.2.1. Participants

A power calculation based on the effect size obtained for preference in Experiment 1 suggested a sample of 649 participants (F = 0.14, α = 0.05, β = 0.90). We recruited 813 participants located in the USA using Amazon’s Mechanical Turk (MTurk). After excluding participants due to invalid responses on an attention-check, and the final sample was N = 724 (62.8% male, 36.7% female, and 0.4% other).

3.2.2. Design, procedure and measures

In Experiment 2, we used the same experimental design as described in Experiment 1, with the difference that the descriptive social norm now promoted the larger pile of toilet paper (the extra-long alternative) rather than the smaller pile (the extra soft alternative). As in Experiment 1, the two main outcome variables were choice and preference.

Participants rated their childhood scarcity on five-point scales ranging from 1 (strongly disagree) to 5 (strongly agree) with three items adapted from Griskevicius et al. (2013): 1) “My family usually had enough money when I was growing up”, 2) “I grew up in a relatively wealthy neighborhood”, and 3) “I felt relatively wealthy compared to the other kids in my school”). We also measured childhood family income. In line with Hansla and Johansson (2020) an index of childhood socioeconomic status (labeled CSES) was constructed by averaging participants’ responses across these four items.

As a measure of control, the current general subjective resources (labeled CRS) of a participant were captured with an index of three statements: “I think that my resources in general are… “(-3 too small, +3 more than enough); “My resources can be described as…” (-3 scarce, +3 abundant) and “…”Currently my resources are…” (-3 small, + 3 big).

3.3. Results

3.3.1. Choice

As may be seen in Tables 3 and 4, in a logistic regression with choice as dependent variable, Model 2.1 shows support for main effects of the descriptive norm in support of H1 and the resource cue in support of H2b. Model 2.2, which includes the resource cue by descriptive norm interaction term, also shows support for H2b, which predicted a main effect of resource cue with the larger pile alternative (here the extra-long option) chosen more often. As previously, no interaction effects were detected.

Table 3 Descriptive statistics for choice (% choosing “extra-long”) in Experiment 2.

Condition	Sample size	Choice
Control	180	31.1%
Descriptive norm	177	49.2%
Scarcity cue	187	42.2%
Descriptive norm and Scarcity cue	180	65%

Table 4 Results from hierarchical logistic regression analyses of participants’ choices of “extra soft” (coded as 0) vs. “extra-long” (coded as 1) toilet paper in Experiments 2.

	Experiment 2
	Model 2.1			Model 2.2
Independent variables	B	SE	p	B	SE	p
Constant	−0.84	0.14	< 0.001	−0.80	0.16	< 0.001
Descriptive norm	0.85	0.15	< 0.001	0.76	0.22	0.001
Scarcity cue	0.57	0.15	< 0.001	0.48	0.22	0.027
Descriptive norm × Scarcity cue				0.17	0.31	0.579
Nagelkerke R^2	0.079			0.079
−2 Log-likelihood	956.6			956.3
N	724

Notes. Bs are unstandardized regression coefficients (logged-odds effects) with associated standard errors. Please note that the descriptive social norm manipulation was aligned with a “large pile” scarcity resource cue. A model with CSES and CRS measure added showed no significant effects whatsoever.

3.3.2. Preference

A factorial 2 × 2 ANOVA with resource cue and descriptive social norm as independent variables on rating of preference, revealed a significant main effect of norm (F(1, 720) = 24.33, p < .001), showing that preferences were stronger in the norm condition (M = 4.31, SD = 1.98) compared to the control condition (M = 3.63, SD = 2.20, d = 0.33, 95% CI [0.12, 0.54]). Furthermore, although the main effect of resource cue was significant (F1, 720, 6.12, p = .014), the 95% confidence interval for the effect size comparing the resource cue condition to the control condition overlapped zero (d = 0.17, −0.04, 0.37). Consequently, the Bonferroni corrected post-hoc test shows a non-significant difference between the resource cue condition and the control condition. Compared to the control condition, no significant effects were found for the resource cue (d = 0.18, 95%CI [0.03, 0.32]), neither were any interaction effects detected (F(1, 720) = 1.14, p = .61) (Table 5).

Table 5 Descriptive statistics for preference in Experiment 2 including sample sizes, means and standard deviations, and effect sizes d comparing each condition to the control condition.

Condition	Sample size	Mean (sd)	Effect size d (95% CI)
Control	180	3.62 (2.20)
Descriptive norm	177	4.31 (1.98)	0.33 (0.12, 0.54)
Scarcity cue	187	3.99 (2.27)	0.17 (-0.04, 0.37)
Descriptive norm and Scarcity cue	180	4.78 (1.97)	0.56 (0.35, 0.77)

3.3.3. Explorative analyses

When a third model including the CSES and CRS measures was added to the regression no significant effects whatsoever were detected.

3.4. Discussion

Experiment 2 replicated the effect of descriptive social norms, showing that the participants’ choices were influenced by what other people choose. Adding to Experiment 1, these results show a consistent influence of the descriptive social norm manipulation. That is, although one choice alternative included a financial motive (extra-long), while the other included a hedonic motive (extra soft), adding a descriptive social norm manipulation promoted both of these choice alternatives respectively.

For the resource-state manipulations, Experiment 2 corroborated the descriptive tendency obtained in Experiment 1 by supporting H2b; predicting a main effect of resource cue with the larger pile alternative (here the extra-long option) chosen more often. Once again, no interaction effects were detected.

Concerning preference, Experiment 2 replicated the effect of Experiment 1, showing that preferences were affected by descriptive social norms, but not by the resource cue. We encourage future research to explore if and how the psychological processes might differ between these social-influence techniques.

4. General discussion

We designed two experiments including 384 and 724 participants respectively, to test the effects of descriptive social norms and resource-state information on consumers’ product choices and preferences. Both experiments showed clear effects of descriptive social norms, demonstrating that participants conformed to other people’s choices. Furthermore, Experiment 1 found a descriptive tendency for choice to be influenced by a resource cue, in how more participants chose the toilet paper from the larger pile, compared to a control condition. Experiment 2 corroborated this tendency effect with statistical significance.

Our results suggest that both descriptive social norms and cues of resource state can influence consumer behavior, and that these influences are additive rather than interactive. Moreover, participants’ preferences for their chosen product were stronger in the descriptive norm compared to the resource cue condition. This effect is in line with previous research (Oruc, 2015).

Although previous research has demonstrated that descriptive social norms affect behavior in consumer settings (e.g., Dorn & Stöckli, 2018; Sparkman & Walton, 2017), none of these studies have examined if and how, when combined, behavioral effects of descriptive norms and cues of resource states are additive, or possibly interact. This study thus adds to prior research by showing that descriptive social norms and resource cues can indeed affect behavior independently and additively. We encourage future research to examine if and how such influences can be generalized across different types of behaviors.

The obtained effect sizes of choice for descriptive social norms in Experiment 2 were larger than expected from past studies (d = .42). One explanation for this effect is that the experimental paradigm used was based on” implicit social norms” (see Bergquist et al., 2019), communicating social norms implicitly rather than using explicit messages. The obtained effect size in Experiment 1 was smaller, however, this experiment had lower statistical power, making random errors more influential. The results from Experiment 1 alone should therefore be interpreted with caution. Although the most valuable conclusions can be drawn from Experiment 1 and 2 in combination, we argue that the findings from Experiment 2 should be given more weight than Experiment 1.

In contrast to the meta-analytic effect of scarcity on purchase intention (Oruc, 2015), we found the resource cue to influence behavior, such that participants chose more from the larger pile. One plausible explanation is that perceived scarcity is dependent on the desirability of the scarce product, therefore showing an effect of scarcity for desirable products but not for undesirable products (Parker & Lehmann, 2011). It should also be noted that the effects of resource cue were here assessed using a shelf-based scarcity research paradigm, while there are other scarcity manipulations such as time-restricted supply, quantity-restricted supply, time-restricted deals, governmental bans, censorship, and age restrictions (see Oruc, 2015), that may or may not have similar effects. Finally, shelf-based scarcity also differs from other quantity limits such as limited edition, because a limited edition product is sure to disappear from the market after it is sold out, making the expected resource depletion permanent instead of temporary (Oruc, 2015).

One limitation of Experiment 1 is that power calculations were made based on the main effect of descriptive norms, whereas a more accurate calculation would have also calculated the expected effect of resource state-information. A power calculation of the expected interaction between descriptive norms and resource cues would have been preferable, yet, to the best of our knowledge, no past study has tested such an interaction. The power calculations in Experiment 2 were therefore based on the obtained effects in Experiment 1.

Another limitation is that both experiments manipulated the “extra-long” as scarce and “extra soft” as abundant, while the descriptive norm promoted the “extra-long” in Experiment 1, and the “extra soft” in Experiment 2. Although this design was confounded with the resource-state manipulations, it implies that we demonstrated the influential effect of descriptive norms even though one alternative is also signaling financial motives (extra-long), while the other is signaling hedonic motives (extra soft). We note, however, that most participants chose the extra soft option in the baseline condition. These results suggest that the influence of descriptive social norms could be generalized, while it limits the generalizability of the resource-cue effect. We encourage future research to look at the effects of relevant consumer motives and examine possible interactions between experienced scarcity and such motives.

Taken together, we consistently found separate effects of descriptive social norms and resource cues, supporting an additive model for product choice. Looking at preference, we found a consistent effect of descriptive social norms causing the consumer to show stronger preference for the chosen product compared to the control condition. Interestingly, no significant effects for preference were observed in the resource cue condition. These results suggest that descriptive norms and resource-state information are additively influential when combined. We encourage future research to examine the processes driving these social influence techniques and their real-world applications.

Ethical statement

This research was conducted in accordance with the ethical standards prescribed by the revised Helsinki Declaration.

Disclosure statement

The authors certify that they have NO affiliations with or involvement in any organization or entity with any financial interest, or non-financial interest, in the subject matter or materials discussed in this manuscript.

Data availability statement

All data is available at https://osf.io/n3ez9/

Additional information

Funding

The Kamprad Family Foundation, grant number 20200135.