A meta-analysis of the pique technique of compliance

Abstract

A random-effects meta-analysis was undertaken to examine the effectiveness of the pique technique of compliance-gaining. Results indicated an overall significant effect of the pique technique on increased compliance rate across 17 comparisons (r = .27). Moderator analyses showed the technique worked significantly better than controls in gaining compliance when a smaller amount was requested, when the reason for the request was included, and when the technique was used in France. In addition, the pique technique was more successful than control requests in terms of the total amount of money or time given by participants and the percentage of participants inquiring about the requests. Theoretical and practical implications for the success of pique technique are discussed for future research.

Keywords:

• Pique technique
• persuasion
• compliance
• meta-analysis

The pique technique as a means of gaining compliance was first studied by Santos, Leve, and Pratkanis in 1994 where confederates acted as panhandlers on the wharf in Santa Cruz, California and asked passersby for money using either a traditional request (‘Can you spare a quarter?’) or an unusual request (‘Can you spare 17 cents?’). The results indicated that a higher percentage of participants gave money when presented with the unusual request than when presented with the traditional request. The authors labeled this unusual request a pique request and it might be said the atypical request piqued the interest of targets on the wharf.

It is important to note that despite increased compliance with the pique request, there was no accompanying difference in the amount of money given by strangers on the wharf when compared to control requests in their seminal analysis. The results also showed that participants in the unusual pique request condition were more likely to ask questions regarding the need for the money (11%) than those in the typical request condition (1%).

Although not tested in the authors’ first field studies, Santos, Leve, and Pratkanis (1994) suggested a theoretical process to account for the pattern of results. It was argued that the pique technique is effective in increasing compliance because the unusual request disrupts the refusal script likely activated when a solicitor asks for money. Most individuals are hardwired not spend time and effort considering a solicitor’s request. It is likely that unwilling targets respond to the solicitor’s request for help with an outright refusal or they divert their attention to indirectly communicate their refusal. When made salient, the pique technique disrupts this heuristic processing by arousing people’s curiosity and focusing their attention on the unusual request.

Since Santos et al.’s (1994) seminal work, several studies have been conducted on the pique technique as a compliance gaining strategy, some designed to replicate the findings and some to examine the boundary conditions for the procedure’s success. For example, in a replication of Santos et al.’s (1994) study, Burger, Hornisher, Martin, Newman, and Pringle (2007, Study 1) reported that the compliance rate was higher in the pique technique condition (42%) than in the control condition (19%) and observed that the amount of money given by willing participants was larger in the pique technique condition ($0.26) than in the control condition ($0.07). Burger et al. (2007) also found that 30% of participants asked the requester about the need for the money in the pique technique condition, whereas no participants inquired about the need for money in the control condition.

Guéguen, Meineri, Pascual, Girandola, and Silone (2015) found that the pique technique was effective in increasing compliance to a time request (compared to the typical monetary request), indicating that the procedure may also work in other compliance-gaining contexts. Moreover, Guéguen and Lamy (2016, Study 2) found that the effect of the pique technique was greater in gaining compliance to a money request when the requester stated an explicit reason asking for the money in the pique technique condition (74%) than when a reason was not given in the pique technique condition (44%). This latter finding would support the notion that target responses to requests for help are mindless and based on scripts (see Langer, Blank, & Chanowitz, 1978).

Although a considerable body of research supports the effectiveness of the pique technique, the results have not been uniform when the tests of the pique technique are conducted in different contexts. Some studies have found little or no effect of the pique technique on the compliance rate, and some have even found a negative effect (e.g., Santos et al., 1994, Experiment 1 & 2, low amount; Feeley, Dietrich, & Musone, 2014, Study 1 & 2). In addition, effect sizes for the compliance rate across studies range from −.08 to .61, using r as the effect size. The reason for the heterogeneity among studies remains unclear, but several potential factors might help explain the inconsistent findings. For example, studies often vary depending on the amount requested in the pique technique conditions (e.g., 17 cents vs. 37 cents vs. 42 cents vs. 3.17 dollars), type of target request (money vs. time), dependent variable (intention to donate vs. actual donation), age of subject (students vs. adults), sex of requester (male only vs. female only vs. both male and female), type of study (field study vs. laboratory study), whether or not the reason for request was included when asking for help, and the country in which the experiments were conducted (United States vs. France). A goal of the current review is to attempt to identify moderators that significantly explain variability among study findings.

The total amount of donations collected from participants have not been uniform across studies. Some studies have found that the pique technique increased the grand amount of donations given by participants (e.g., Burger et al., 2007, Study 1; Guéguen & Lamy, 2016, Study 2, justification), but other studies have found that there was no difference between the pique technique and control conditions for the amount of donation given by participants (e.g., Santos et al., 1994, Experiment 1, low & high amounts; Guéguen, Meineri, Pascual, & Girandola, 2015). Moreover, the results of the percentage of participants asking the requester about the need for the money were also different across studies. Some studies have found that the percentage of participants inquiring about the need for the money was higher in the pique technique condition than in the control condition (e.g., Burger et al., 2007, Study 1 & 2; Santos et al., 1994, Experiment 1 & 2), whereas other studies have found that there was no difference between the two request conditions (e.g., Guéguen & Lamy, 2016, Study 1 & 2, no justification). These and other potential factors will be investigated in the current paper.

In an attempt to explain the varied results in relation to studying the effectiveness of pique technique, a meta-analysis will be conducted of the research literature in this area of inquiry. Specifically, the purpose of the current analysis is threefold: (a) to test the overall effect size of the pique technique for compliance rate and to identify potential moderators to account for variance in effects across studies, (b) to examine the ultimate utility of the pique technique by testing the difference in the amount of donation given by participants between the pique technique and control conditions, and (c) to demonstrate why the pique technique works by investigating the difference in the percentage of participants inquiring about the requests between the pique technique and control conditions. If the pique technique can be explained by a reduction or halt in the target’s refusal script, then it is to be expected a greater number of inquiries about the need for the donation will be associated with pique requests. The next section will describe the meta-analytic procedures to address the study purposes.

Method

Search strategies and inclusion criteria

Two methods were used to search the research literature on the pique technique with journal articles, book chapters, dissertations, and conference papers reviewed for study inclusion. First, a search was conducted in the databases Communication & Mass Media Complete, PsycINFO, PubMed, Social Science Citation Index, Google Scholar, and ProQuest Dissertation and Theses for the 1994−2016 time period, using the search terms ‘pique technique,’ ‘unusual request,’ ‘can you spare 17 cents,’ and ‘can you spare 37 cents.’ Second, we performed an ancestry search of relevant studies by reviewing reference sections of studies found through the first method (see Cooper, 1989).

All studies identified were carefully considered for review and for a study to be included, it had to meet the following inclusion criteria: each study had to (1) have a pique technique condition that asked participants for unusual or strange amounts of money or time (e.g., 17 cents, 37 seconds), (2) have a control condition that asked participants for typical amounts of money or time (e.g., a quarter, little time), (3) include one or more measures of outcomes such as compliance rate, mean amount of donation per donor, total amount of donation, and percentage of participants making inquiry about request, and (4) report sufficient statistical information to calculate effect sizes across relevant studies.

This procedure yielded a set of six articles, which reported the results of 17 comparisons for the compliance rate (N = 2,136), 16 comparisons for the mean amount of donation per donor and the total amount of donation (N = 1,836), and 10 comparisons for the percentage of participants making inquiry about request (N = 1,481).

Coding for moderators

Systematic differences in contextual and methodological characteristics of studies may cause variation in reported effects (Bijmolt & Pieters, 2001; Sutton, Abrams, Jones, Sheldon, & Song, 2000). Thus, we examined eight potential moderators related to contextual and methodological factors. The contextual factors are primarily related to the characteristics of request messages and the methodological factors are related to the characteristics of participants and dependent variables. The moderators considered are defined in the next subsection (see Table 1).

Table 1. Effect sizes and moderators associated with the compliance rate.

Study	Pique		Control		r	a	b	c	d	e	f	g	h
	%	Complaint subjects/n	%	Complaint subjects/n
Burger et al. (2007, Study 1, combined pique conditions)	41.7	85/204	18.8	22/117	.233	2	2	1	2	2	4	1	1
Burger et al. (2007, Study 2, control 25, combined pique conditions)	42.1	64/152	15.0	12/80	.275	2	2	1	2	2	4	1	1
Feeley et al. (2014, Study 1)	15.4	4/26	21.4	6/28	−.078	4	2	1	2	2	2	1	1
Feeley et al. (2014, Study 2)	21.4	6/28	25.0	7/28	−.042	4	2	1	2	2	2	1	1
Feeley et al. (2014, Study 3)	21.4	9/42	16.3	7/43	.066	4	2	1	2	2	3	1	1
Guéguen, Meineri, Pascual, and Girandola (2015)	45.0	18/40	15.0	6/40	.327	2	2	1	2	2	2	1	2
Guéguen, Meineri, Pascual, and Girandola (2015, reframe)	57.5	23/40	20.0	8/40	.385	2	2	1	2	2	2	1	2
Guéguen, Meineri, Pascual, Girandola, and Silone (2015)	60.0	90/150	42.0	63/150	.180	2	2	2	2	2	2	1	2
Guéguen and Lamy (2016, Study 1)	72.5	29/40	12.5	5/40	.607	3	2	1	2	2	2	1	2
Guéguen and Lamy (2016, Study 2, justification)	74.0	37/50	30.0	15/50	.440	2	1	1	2	2	1	1	2
Guéguen and Lamy (2016, Study 2, no justification)	44.0	22/50	14.0	7/50	.331	2	2	1	2	2	1	1	2
Guéguen and Lamy (2016, Study 3, final sum mention)	88.0	44/50	40.0	20/50	.500	3	1	1	2	2	2	1	2
Guéguen and Lamy (2016, Study 3, no final sum mention)	58.0	29/50	26.0	13/50	.324	3	1	1	2	2	2	1	2
Santos et al. (1994, Experiment 1, high amount)	30.6	22/72	15.3	11/72	.182	2	2	1	2	2	2	1	1
Santos et al. (1994, Experiment 1, low amount)	42.5	31/73	30.6	22/72	.124	1	2	1	2	2	2	1	1
Santos et al. (1994, Experiment 2, high amount)	75.0	30/40	42.5	17/40	.330	2	2	1	1	1	4	2	1
Santos et al. (1994, Experiment 2, low amount)	74.4	29/39	65.0	26/40	.102	1	2	1	1	1	4	2	1

Notes: a – Request amount in pique technique condition (1 = 17 cents, 2 = 37 cents or seconds, 3 = 42 cents, 4 = 3.17 dollars); b – Reason for request (1 = Yes, 2 = No); c – Target request (1 = Money, 2 = Time); d – Dependent variable (1 = Intention to donate money or time, 2 = Actual donation of money or time); e – Age of subject (1 = Students, 2 = Adults); f – Sex of requester (1 = Male only, 2 = Female only, 3 = Both male and female, 4 = NA); g – Type of study (1 = Field study, 2 = Laboratory study); h – Country (1 = United States, 2 = France).

Request amount

Studies were coded to indicate whether the amount of money or time requested in pique technique condition. As it turns out only four different amounts have been used in studies – 17 cents, 37 cents (or seconds), 42 cents, and 3.17 dollars.

Reason for request

Studies were coded to indicate whether the reason for asking for the money or time was included in the request messages (Yes/No). For example, Guéguen and Lamy (2016, Study 2) included the following reason for asking for money in the request messages: ‘Would you happen to have [some change] or [37 cents] so that I could buy a sandwich, please? (p. 119).’

Target request

Each field experiment was identified as requesting money or time (e.g., sparing time to answer a survey) from strangers.

Dependent variable

The type of dependent variable was coded as either intention to donate money or time or actual donation of money or time.

Age of subject

Papers were coded for whether targets were students who were approached on a college campus or adults who were approached from the general public and it is assumed college samples are younger in age than general samples.

Sex of requester

It was noted if the sex of requesters was male only, female only, or both male and female.

Type of study

Studies were coded based on whether they were conducted in a field or using a laboratory or controlled setting.

Country

Studies were coded to indicate whether the country in which experiments were conducted was the United States or France.

Calculation of effect sizes

Effect sizes for the compliance rate of all studies were calculated in the metric of r for ease of interpretation using formulas provided by Lipsey and Wilson (2001).¹ The effect size indicates the difference between the pique technique and the control conditions; positive effects represent an advantage of the pique technique condition. The effect sizes were calculated based on cell frequencies of a 2 by 2 contingency table with the pique technique condition and control condition as categorical variables with two levels (yes or no for compliance) each.

To derive effects from each study, two general rules guided analyses. First, when a study employed an experimental design comparing a pique technique condition with a corresponding control condition with multiple treatment arms, each effect contributed uniquely to the meta-analysis rather than collapsing the treatment arms together. For example, Santos et al. (1994, Experiment 1) employed a 2 (pique technique vs. control) by 2 (low amount vs. high amount) between-subjects design. For the current analysis, two independent effects (i.e., comparisons between the pique and the control condition in both the low amount and high amount conditions) were included. Second, when a study included multiple control conditions in which different request amounts were employed, the control condition asking for the amount (i.e., a quarter or any change) was compared with the corresponding pique technique amount in terms of amount (see Santos et al., 1994). Thus, while the multiple treatment arms are compared to the same control group, no single target individual is represented in more than one experimental condition.

Analysis

Analyses were conducted using a random-effect model based on DerSimonian and Laird (1986).² Each analysis was performed in the metric of Fisher’s z, with weighted averages of effect sizes and 95% confidence intervals (CIs). Heterogeneity tests were conducted using I² and Q statistics. The I² test addresses the degree of variance between studies, and indicates 25, 50, and 75% correspond to low, moderate, and high between-group variance, respectively (Higgins, Thompson, Deeks, & Altman, 2003).

Moderator analyses were conducted with mixed-effects models. Q_B, a measure of significance for the moderators, was computed under the mixed-effects model. The effect size, variance, and 95% CI for each subgroup were also calculated under the mixed-effects model. All computations were computed using the statistical software package Comprehensive Meta-Analysis 2.2 (Borenstein, Hedges, Higgins, & Rothstein, 2006).

Results

The results are presented in three sections. The first section presents results for the compliance rate, the second section presents results for the donation amount such as the mean amount of donation per donor and total amount of donation, and the third section presents results for the percentage of subjects making an inquiry about request. Before conducting the analyses, we examined a standard funnel plot to test for evidence of publication bias. A serious asymmetry would be expected in the funnel plot when publication bias may exist. As shown in Figure 1, the funnel plot did not show serious asymmetry. Thus, there is little concern for publication bias in the data. This lack of publication bias was further confirmed by both Begg and Mazumdar’s (1994) rank correlation test (p = .68) and Egger, Smith, Schneider, and Minder’s (1997) regression test (p = .67).

Figure 1. Funnel plot of standard error by Fisher’s Z.

Compliance rate

Overall effect

The overall weighted average correlation under the random-effect model for the 17 effects associated with compliance rate was r = .27, z = 6.54, p < .001, 95% CI = .19, .34, N = 2,136. The finding indicated that the pique technique conditions yielded a significantly higher compliance rate than did the control conditions. The homogeneity test (Q_T [16] = 52.70, p < .001, I² = 69.6%) represented a moderate amount of heterogeneity. Thus, it appears moderator analyses are warranted.

Moderator effects

Table 2 provides a summary of the results concerning the effects of the eight study moderators, considered individually. The results indicated that request amount (Q_B [3] = 23.40, p < .001), reason for request (Q_B [1] = 8.44, p < .01), and country (Q_B [1] = 10.30, p < .001) were significant moderators. Specifically, the pique technique worked primarily when lesser amounts were requested in the pique technique conditions, indicating that the effects for the small amounts of request were significantly positive (17 cents, r = .12, k = 2, p < .10; 37 cents, r = .27, k = 9, p < .001; 42 cents, r = .48, k = 3, p < .001), whereas the effect for the large amount of request did not differ significantly from zero (3.19 dollars, r = −.004, k = 3, p = .953). For the subgroups of studies in which the small amounts were solicited in the pique conditions (17 cents, 37 cents, and 42 cents), effect sizes significantly increased when the amounts of request increased.

Table 2. Moderator analyses for the compliance rate.

Moderator [QBetween]	k	r	95% CI	z	p	QWithin
Overall	17	.265	[.19, .34]	6.54	.000	52.70^***
Request amount in pique technique condition [Q (3) = 23.40, p = .000]
17 cents	2	.116	[-.02, .25]	1.73	.084	0.03
37 cents or seconds	9	.274	[.22, .33]	8.91	.000	10.41
42 cents	3	.482	[.31, .62]	4.99	.000	5.97
3.17 dollars	3	−.004	[-.15, .14]	−0.06	.953	0.76
Reason for request [Q (1) = 8.44, p = .004]
Yes	3	.424	[.32, .52]	7.26	.000	2.26
No	14	.229	[.15, .31]	5.46	.000	38.26^***
Target request [Q (1) = 1.72, p = .189]
Money	16	.271	[.19, .35]	6.20	.000	50.44^***
Time	1	.180	[.07, .29]	3.14	.002	0.00
Dependent variable [Q (1) = 0.18, p = .674]
Intention to donate money or time	2	.219	[-.01, .43]	1.85	.064	2.21
Actual donation of money or time	15	.270	[.19, .35]	6.16	.000	50.23^***
Age of subject [Q (1) = 0.18, p = .674]
Students	2	.219	[-.01, .43]	1.85	.064	2.21
Adults	15	.270	[.19, .35]	6.16	.000	50.23^***
Sex of requester [Q (3) = 7.42, p = .060]
Male only	2	.387	[.26, .50]	5.68	.000	0.80
Female only	10	.268	[.14, .39]	4.07	.000	41.89^***
Both male and female	1	.066	[-.15, .28]	0.60	.550	0.00
NA	4	.244	[.17, .31]	6.59	.000	2.61
Type of study [Q (1) = 0.18, p = .674]
Field study	15	.270	[.19, .35]	6.16	.000	50.23^***
Laboratory study	2	.219	[-.01, .43]	1.85	.064	2.21
Country [Q (1) = 10.30, p = .001]
United States	9	.165	[.09, .24]	4.15	.000	13.41
France	8	.386	[.28, .49]	6.44	.000	23.50^**

^** p < .01

^*** p < .001.

In addition, effect sizes were significantly larger when the reason for request was included in the request messages (r = .42, k = 3, p < .001) as opposed to when the reason for request was not included (r = .23, k = 14, p < .001), and when the experiments were conducted in France (r = .39, k = 8, p < .001) as opposed to when the experiments were conducted in the United States (r = .17, k = 9, p < .001). Meanwhile, the other study moderators considered in the current meta-analysis (target request, dependent variable, age of subject, sex of requester, and type of study) did not significantly account for the variation in effects across studies. As you can see in the tables, the number of comparisons for some moderator factors was as little as 1, so caution is warranted in interpreting study findings.

Donation amount

A second set of analyses was undertaken to determine if the mean amount of donation per donor and the total amount of donation generated in pique technique conditions were significantly different from control conditions. To conduct these analyses, we calculated the mean amount of donation per donor and the total amount of donation for both pique technique and control conditions from 16 comparisons (see Table 3). Because effect sizes for the mean amount of donation per donor and the total amount of donation could not be calculated in our observations of these data by the lack of statistical information, it was not permissible to use traditional meta-analytic techniques to test for differences. Instead, we used a Wilcoxon Signed Rank Test to examine the difference in the mean amount of donation per donor and the total amount of donation between the pique technique and the control conditions.

Table 3. Mean amount of donation per donor and total amount of donation.

Study	Mean donation				Total donation
	Pique		Control		Pique	Control
	Mean	Complaint subjects/n	Mean	Complaint subjects/n	Total	Total
Burger et al. (2007, Study 1, combined pique conditions)	$0.45	85/204	$0.37	22/117	$37.18	$8.19
Burger et al. (2007, Study 2, control 25, combined pique conditions)	$0.48	64/152	$0.27	12/80	$30.51	$3.20
Feeley et al. (2014, Study 1)	$4.00	4/26	$6.17	6/28	$16.0	$37.0
Feeley et al. (2014, Study 2)	$4.33	6/28	$3.14	7/28	$26.0	$22.0
Feeley et al. (2014, Study 3)	$3.72	9/42	$3.43	7/43	$33.5	$24.0
Guéguen, Meineri, Pascual, and Girandola (2015)	€0.16	18/40	€0.13	6/40	€2.88	€0.78
Guéguen, Meineri, Pascual, and Girandola (2015, reframe)	€0.37	23/40	€0.19	8/40	€8.51	€1.52
Guéguen and Lamy (2016, Study 1)	€0.46	29/40	€0.36	5/40	€13.34	€1.80
Guéguen and Lamy (2016, Study 2, justification)	€0.61	37/50	€0.39	15/50	€22.57	€5.85
Guéguen and Lamy (2016, Study 2, no justification)	€0.49	22/50	€0.37	7/50	€10.78	€2.59
Guéguen and Lamy (2016, Study 3, final sum mention)	€0.47	44/50	€0.42	20/50	€20.68	€8.40
Guéguen and Lamy (2016, Study 3, no final sum mention)	€0.47	29/50	€0.38	13/50	€13.63	€4.94
Santos et al. (1994, Experiment 1, high amount)	$0.41	22/72	$0.63	11/72	$9.02	$6.93
Santos et al. (1994, Experiment 1, low amount)	$0.32	31/73	$0.41	22/72	$9.92	$9.02
Santos et al. (1994, Experiment 2, high amount)	$0.30	30/40	$0.35	17/40	$9.00	$5.95
Santos et al. (1994, Experiment 2, low amount)	$0.41	29/39	$0.31	26/40	$11.89	$8.06
Overall	1.09	482/996	1.08	204/840	17.21	9.39

For the mean amount of donation per donor, the results indicated that there was no significant difference between the pique technique conditions (M = 1.09, SD = 1.46) and the control conditions (M = 1.08, SD = 1.69), z = 1.63, p = .103, r = .29. Meanwhile, the results for the total amount of donation showed that the pique technique conditions (M = 17.21, SD = 10.09) produced a significantly larger total amount of donation than did the control conditions (M = 9.39, SD = 9.89), z = 2.79, p < .01, r = .49.

Percentage of making inquiry about request

Lastly, we conducted an analysis to determine if the pique technique conditions resulted in a greater number of questions about requests than did control conditions. To conduct this analysis, we calculated the percentage of participants making inquiry about request for both pique technique and control conditions from 10 comparisons (see Table 4). As the case in amount of donations, because effect sizes for the percentage of participants making inquiries about request amount could not be computed in our observations of these data by the lack of statistical information, a Wilcoxon Signed Rank Test was used to examine the difference in the percentage of participants making inquiry about request between the pique technique and control conditions. The results indicated that participants in the pique conditions (M = 22.2, SD = 20.3) asked significantly more questions about request than those in the control conditions (M = 6.2, SD = 11.2), z = 2.52, p < .05, r = .56.

Table 4. Percentage of participants asking about the request.

Study	Inquiry about request
	Pique		Control
	%	Inquiring subjects/n	%	Inquiring subjects/n
Burger et al. (2007, Study 1, combined pique conditions)	29.9	61/204	0.0	0/117
Burger et al. (2007, Study 2, control 25, combined pique conditions)	37.5	57/152	0.0	0/80
Feeley et al. (2014, Study 2)	18.0	5/28	7.0	2/28
Guéguen, Meineri, Pascual, Girandola, and Silone (2015)	8.0	12/150	1.3	2/150
Guéguen and Lamy (2016, Study 1)	0.0	0/40	0.0	0/40
Guéguen and Lamy (2016, Study 2, no justification)	0.0	0/50	0.0	0/50
Santos et al. (1994, Experiment 1, high amount)	13.9	10/72	0.0	0/72
Santos et al. (1994, Experiment 1, low amount)	8.2	6/73	1.4	1/72
Santos et al. (1994, Experiment 2, high amount)	57.5	23/40	20.0	8/40
Santos et al. (1994, Experiment 2, low amount)	48.7	19/39	32.5	13/40
Overall	22.6	188/820	6.1	24/661

Discussion

The purpose of the current study was to examine the magnitude of the pique technique effect for the compliance rate and to investigate if study moderators explain the heterogeneity surrounding the pique technique studies. Moreover, this study aimed at testing if there is the difference between the pique technique and control conditions for the total amount of donations given by participants and the percentage of participants asking about the requests. The following sections summarize study results and their implications for compliance-gaining research.

Overall effect of the pique technique

The present review summarized the effect sizes for the compliance rate from 17 comparisons associated with the pique technique and found a significant overall mean effect size of r = .27. This effect size was larger than the meta-analytic estimates of the effect sizes for other common compliance gaining techniques including the foot-in-the-door r = .11 (Dillard, Hunter, & Burgoon, 1984), the door-in-the-face r = .13 (Feeley, Anker, & Aloe, 2012), and the legitimization of paltry favors r = .17 (Lee, Moon, & Feeley, 2016). Thus, an effect of this magnitude suggests that the pique technique is an effective message strategy in gaining compliance, and suggests that from a practical standpoint, charitable organizations may benefit from utilizing the pique technique when eliciting charitable giving.

Moderators

The pattern of effect sizes across studies indicated a fair amount of heterogeneity and this realization puts the spotlight on factors that may help to explain this state of affairs. Three moderator variables―request amount, reason for request, and country―play significant roles in influencing the pique technique effects.

With respect to the request amount, the current study found that the increase of request amount resulted in the increase of compliance when the small amounts of request (17 cents, 37 cents, and 42 cents) were solicited in the pique conditions, whereas the pique technique was not effective when the large amount of request (3.17 dollars) was asked for in the pique conditions. To provide a more fine-grained analysis of this finding, we additionally examined the effect sizes by the request amounts in the pique technique and control conditions. Of the total 17 control comparisons, 3 comparisons requested ‘one quarter,’ 11 asked for ‘some change,’ and the remaining 3 comparisons solicited ‘3 dollars.’ The effect size for 37 cents (r = .28, k = 1) was larger than that for 17 cents (r = .12, k = 2) when a quarter was requested in the control conditions. In addition, the effect size for 42 cents (r = .48, k = 3) was larger than that for 37 cents (r = .28, k = 8) when ‘some change’ was requested in the control conditions. However, it was not permissible to compare the effect size for 3.17 dollars with those for other amounts when 3 dollars was requested in the control conditions because this request was only compared to 3.17 dollars in the three comparisons (Feeley et al., 2014). These results indicated that in the cases in which the small amounts of request were solicited, when requests of larger amounts were solicited in the pique conditions, compliance tends to increase regardless of the amounts of the requests in the control conditions. Also worth noting, the current results of the request amount were inconsistent with those reported by Santos et al. (1994) who found there was no difference in compliance rates due to the size of request amounts in the pique technique conditions.

This last finding with respect to the relationship between effect size and request amount is consistent with Feeley et al.’s (2012) finding with DITF findings; these authors found the procedure works better when compliance is more difficult to come by. If we consider 17, 37, and 42 cents equally unique as request amounts, then it appears the more you ask, the more the technique matters in terms of compliance when compared to asking for conventional amounts of money or time. When being asked to help a person or cause, there are a number of factors that potentially influence one’s decision and it may be the case the active agent in the pique technique is more influential as compliance becomes less likely or, stated differently, when compliance is more costly. The active agent in the pique technique is an important consideration that will be taken up shortly in this discussion section.

Study findings indicated when a reason for the request was provided, the technique worked significantly better. This finding is consistent with Guéguen and Lamy’s (2016) argument that if there is a reason asking for the money, the compliance with the request is increased, particularly when an unusual amount of money is requested. Guéguen and Lamy (2016) claimed that the cause of the increase in the effectiveness of the pique technique, when an explicit reason accompanied, is probably because the request is perceived as more legitimate, demonstrating that the effectiveness of the pique technique increases more when the requester gives a reason asking for the money in the pique technique condition than when a reason is not stated in the pique technique condition. Moreover, our finding of the reason for request appears to be associated with the result reported by Burger et al. (2007) that the pique technique is only effective with participants who ask the requester what the money is for and are told the reason for the need of money from the requester. That is, when an unusual pique request is followed by a reason for the request, more compliance can be obtained because the reason increases the legitimacy of the request that makes it more difficult for people to find a reason to refuse to help the requester.

The present research also found that country was a significant moderator to account for variation in effect sizes across studies on the pique technique. Specifically, the pique technique was effective for both the United States and France, but the mean effect size was greater when experiments were conducted in France than in the United States. In other words, these results imply that the pique technique is a robust persuasive message strategy to elicit compliance across nations, and particularly, the pique technique may be more successful when the pique technique is used in France.

Beyond examining overall effect size of the pique technique for compliance rate and moderators to explain variances in effects across studies, a secondary purpose of this review was to test the ultimate utility of the pique technique. The results of the current study indicated that there was no significant difference in the mean amount of donation per donor between the pique technique and control conditions. Although our results demonstrate that the pique technique conditions do not elicit a greater mean amount of donation per donor compared with control conditions, some people might argue that the pique technique is still beneficial compared with control groups. This is because, even though there was no significant difference between the pique technique and control conditions for the mean amount of donation per donor, the pique technique conditions yielded a significant higher compliance rate than did control conditions. Given this pattern of results, thus, it may be expected that the pique technique condition is superior to the control condition in terms of the total amount of donation which is the result by the multiplication of the compliance rate and the mean amount of donation per donor. In the present research, we tested this notion by examining whether the total amount of donation was different for the pique technique conditions compared with the control conditions. Our results indicated that the pique technique conditions led to a reliably greater total amount of donation than did the control conditions. Thus, it appears that the pique technique has practical utility for requesters in fund-raising drives.

The final goal of the current study was to demonstrate why the pique technique works. The results of the present research indicated that participants in the pique technique conditions asked significantly more questions about the need for money or time solicited by the requester than did those in the control conditions. These results support Santos et al.’s (1994) argument that the pique technique is effective because it disrupts the refusal script commonly generated when a solicitor makes a request by arousing individuals’ curiosity and focusing their attention on the unusual request. That is, the increased curiosity and attention on the unusual request might contribute to the increased compliance.

Toward a theory of pique findings

The pique technique works by halting the typical mindless refusal script enacted by strangers when asked for help on the wharf, in the hallways, or in a shopping mall. The unique request amount causes targets to pause and take notice of what is being asked and perhaps why it is needed, rather than mindlessly declining to help or refusing to respond altogether. The more mindful target confronted with an unorthodox request amount may take pause and consider the specific need or plight for the individual or charity who has worked out the exact amount needed per willing donor.

This increase in mindfulness, however, does not explain why compliance increases when a reason for the request is provided (Guéguen & Lamy, 2016; Study 2 justification). In the original Langer et al. (1978) experiments, when confederates provided a reason why they needed to cut in line, compliance increased, no matter how placebic the reason (e.g., ‘I need to make copies’), when small favors were sought (i.e., fewer copies needed/less time). However, this would indicate targets are still mindlessly reacting to the request. Providing a reason fulfills the mindless script while asking for an unexpected amount would halt this same script according to original thinking on how pique operates (Santos et al., 1994). Thus, it is unclear the exact role of mindlessness in pique compliance-gaining field experiments.

Future research would go far to consider the pique technique in relation to other popularly studied techniques for gaining compliance in communication. For example, the legitimization of paltry favors (Lee et al., 2016) increases compliance by stressing the need for compliance and making salient even a few pennies or perhaps 17 cents would help the person or cause. The disrupt-then-reframe, halts the receivers’ refusal script then reframes the asking amount in a more elementary way (e.g., ‘we are asking for $2.00, that’s 200 pennies’). It may be useful to combine two or more of these procedures to get a relatively larger return on investment when seeking compliance from strangers.

It is imperative future research focus on testing the active agent in operation when using the pique technique. As has been learned from FITD research (see Burger’s incisive review), there may be more than one theoretical mechanisms actively in play when receivers are confronted with pique requests so research could consider this possibility.

Limitations

The current review has a limitation to small samples. Although several studies have conducted a meta-analysis on compliance gaining techniques with a comparable number of studies, for example the disrupt-then-reframe k = 14 (Carpenter & Boster, 2009), the legitimization of paltry favors k = 19 (Andrews, Carpenter, Shaw, & Boster, 2008), and the low-ball k = 19 (Burger & Caputo, 2015), the number of comparisons included in the current meta-analysis (k = 17) is small. When the number of comparisons is small, statistical power to detect significant differences can be decreased (Lipsey & Wilson, 2001).

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1. According to Lipsey and Wilson (2001, p. 201), the following formula was used to compute effect sizes based on a 2 × 2 table with cell frequencies of a, b, c, and d.

In instances where the total sample size and the χ² statistic with df = 1 was available, an alternative formula was also used:

2. The statistical software package used in the current meta-analysis, Comprehensive Meta-Analysis, was programmed based on DerSimonian and Laird’s (1986) procedure.