The Decoy Effect on Choosing Branded and Non-Branded Alcohol-related Products

Dublin Core

Title

The Decoy Effect on Choosing Branded and Non-Branded Alcohol-related Products

Creator

Wang Li

Date

2017

Description

The decoy effect describes a phenomenon that the introduction of a third choice, usually an asymmetrically dominated one, would change the distribution of people’s preferences for the original two options. Monk et al. (2016) found a basic decoy effect on alcohol purchasing decisions. Extending this, the current study examined the impact of the decoy effect on alcohol-related purchasing decisions and whether the addition of brand names would further impact this. A total of 106 participants were asked to make decisions amongst four types of branded and unbranded drinks by completing an online questionnaire. They also completed the AUDIT, assessing problem drinking patterns, and a measure of trait effortful control. Results showed that the decoy appeared to affect alcoholic relative to non-alcoholic drinks, and affected branded products more than non-branded products. The results suggest that the decoy effect might affect alcohol-related purchasing decisions differently to non-alcoholic purchasing decisions, which might have managerial implications for marketers and health implications for hazardous alcohol consumptions.

Subject

alcohol purchasing
decoy effect

Source

All the stimuli were pictures of bottles presented with text descriptions including the number of bottles and the total prices. Participants viewed a total of 80 deals, with 40 alcoholic products and 40 non-alcoholic products. Half of the products were branded and half unbranded. With regards to branded stimuli, pictures of bottles similar to those presented in supermarkets were shown (see Figure 2), whilst unbranded ones used similar unbranded bottles in terms of colour and shape. This allowed the products to correspond with both the diversity of goods in real supermarket but also to avoid unnecessary brand association (e.g. red glass bottles always remind consumers of Coca-Cola; Underwood, 2003).
In the control condition, participants were shown products with two options, one with less bottles but cheaper, and the other with a greater quantity of bottles but more expensive. As such, option 1 represented the competitor option, which was cost-effective, and option 2 represented the target option, which was moderately cost-

effective. The order of cost-effective and moderate-cost effective products was randomised throughout the experiment. In comparison, the experimental conditions also included a decoy option, which presented a product that was the least cost-effective. Although the decoy option itself was unlikely to be selected, it was expected to result in a different distribution of selections from the output of the control group. The sequences of the choice A (cost-effective), B (moderate cost-effective), and the C (decoy) were set randomly.
Alcohol Use Disorders Identification Test (AUDIT). This questionnaire was made up of 10-items concerning personal drinking habits, drinking frequency and amount (Saunders et al., 1993). On a scale of one to over 30, articipants responsed to the questions such as “How many units of alcohol do you drink on a typical day when you are drinking?” and a total AUDIT score was computed. The scores of each question were accumulated and coded. The output of the AUDIT test showed great reliability of this study, M = 15.09, SD = 4.60, Cronbach’s α = .79. It should be noted that although the figure for the AUDIT test was way above eight which was the hazardous cut-off, indicating a possible harmful alcohol use (Babor et al., 2001), this was in line with student’s drinking cultures in UK.
Adult Temperament Questionnaire (ATQ). This questionnaire assessed self-reported effortful control (c.f., Evans & Rothbart, 2007) and comprised of 34-items, such as “When interrupted or distracted, I usually can easily shift my attention back to whatever I was doing before”. Participants were asked to answer the questions by selecting a 7-point Likert Scale (1 = extremely untrue, 7 = extremely true). Their responses were recorded and coded. A small amount of missing data (caused by unexpected errors on the web) was replaced by the mean of the sample on that specific item. The result of the ATQ test revealed internal consistency as well, M = 146.34, SD = 22.82, Cronbach’s α = .85.
Design and Procedure
This study conducted a 4 Stimuli (Beer, Cider, Orange Juice and Water) x 2 Brading (Branded vs unbranded) x 2 Selection (Cost-effectiveness vs. Moderate Cost-effectiveness) within-subjects research design, to examine the possible shifts of selections with the addition of the decoy. Participants were instructed to look at online supermarket choice sets and asked to make a choice out of two (control condition) of three product options (the experimental condition, with the decoy product added). At the beginning, they were asked to imagine that they were in a real supermarket, and they were told that their selections would be dependent on their own preferences. No other information was provided either in oral or on the screen in order to prevent demand characteristics.
The main questionnaire had 80 questions, consisting of 80 trials of stimuli (i.e., 20 trials for beer, water, orange juice, and cider). Also 40 groups of bottles were branded and the other 40 were non-branded. The main questionnaire comprised four web pages, with 20 questions in each page and took approximately 30 minutes to complete. Participants were allowed to take short breaks when they finished one page of questions. There was no time limit for each of the questions as the pressure caused by time constraints has been found to affect one’s decision-making process (Dhar & Nowlis, 1999). Subsequently, participants completed a self-report measure of hazardous drinking behaviour (AUDIT) and the effortful control scale (ATQ). These questionnaires were completed at the end of the experiment to make sure the alcohol-related behaviours were not primed (Monk et al., 2016). At the end of the experiment, participants completed a manipulation check to ensure that they were able to accurately distinguish the cost-effectiveness and the quantity of the products set and that they fully understood the requirements of this study. They were then asked to report if they had consumed alcohol on the day of testing, as alcohol consumption has been shown to affect decision-making and may therefore affect the findings of the experiment (Steele & Josephs, 1990). Therefore, participants who had consumed alcohol before participating in the test were excluded when analysing the decoy effect (n = 8). Finally, participants were fully debriefed after they had finished the whole experiment, and were informed about the true aims and hypotheses of the study.

Publisher

Lancaster University

Format

data/ xlsx

Identifier

Li2017

Contributor

John Towse

Rights

Open

Language

English

Type

Data

Coverage

LA1 4YF

LUSTRE

Supervisor

Charlotte Pennington

Project Level

MSc

Topic

Psychology of Advertising

Sample Size

106? participants were recruited. Thirty of them were male participants and 70 were females

Statistical Analysis Type

ANOVA

Files

Collection

Citation

Wang Li, “The Decoy Effect on Choosing Branded and Non-Branded Alcohol-related Products,” LUSTRE, accessed March 28, 2024, https://www.johnntowse.com/LUSTRE/items/show/18.