Manipulation of independent variables within the stimuli

To test the effect of the independent variables, eight different stimuli were created, as shown in Table 1. Since real, existing products differ greatly on multiple design elements, it was decided to design the product stimuli from scratch with a fictional product such that all design elements, next to the manipulations, could be kept constant for internal validity (Trochim, Donnelly, & Arora, 2015). The product used in the stimuli was from the personal care industry, since research showed that particularly in highly competitive markets where consumers have plentiful choice, effectively designing packaging to signal desired attributes, such as sustainability, is crucial (Hellström & Saghir, 2007; Rundh, 2005; Vila, 2006), specifically since many purchase decisions are made at the point of sale (Clement, 2007; Vila, 2006). This is predominantly the case for fast-moving consumer goods (FMCG) markets, such as the food and personal care industry. Within both these industries, an increasing number of packaging alternatives are offered, including an increasing number of sustainable packages (Adetayo, Opele, & Olasehinde, 2019). For instance, grapes are offered in both plastic and

37 paper packages, but also soap or shampoo is now offered in paper-based packages next to the regular plastic packages. Since most existing research regarding consumer responses to sustainable packaging has been done for food products, and multiple studies request further insights into products from other industries (Marckhgott & Kamleitner, 2019; Ye et al., 2019), a product from the personal care industry was chosen for this study. Within the personal care industry, an evolution in sustainable packaging is seen (e.g. more biodegradable materials) and the importance to successfully communicate desired product features is stressed (Rosette, Yablonski, Mancuso, & Kale, 2012). Therefore, the personal care industry is a very relevant industry to study. The specific product used in the stimuli is a soap bar, since currently both a paper and plastic package material are quite common for a soap bar.

The stimuli were designed in cooperation with a professional agency through Photoshop using 3D mock-up images, which were visually manipulated according to the three independent variables. First, packaging material was manipulated by using two different mock-up images of both a paper and a plastic soap package. Additional visual adjustments were added, such as a zig-zag edge, to make the plastic package appear more like plastic.

Furthermore, surface varnishing was manipulated by making the package’s appearance rather matte or glossy by removing any light reflections in the matte condition and adding shiny light reflections in the glossy conditions. Lastly, environmental claim was manipulated by adding a claim ‘package made of recycled paper’ or ‘package made of recycled plastic’ on the package in the conditions where a claim was present, whereas no information was provided in the no claim conditions. This particular claim was chosen since research showed that unsubstantiated, simple ecological claims are relied upon more by consumers and can increase sustainability perceptions (Chan & Lau, 2004; Gleim et al., 2013; Söderlund & Mattsson, 2020). Moreover, the material being recycled is quite common for both paper and plastic, which would thus not cause any confounding effects. As mentioned, all stimuli varied simply on these three

38 manipulated variables, while all other design elements were kept constant across the stimuli to avoid confounding effects and be able to detect causation (Shadish et al., 2002). Furthermore, to control for the influence of familiarity with the brand or product (Orth, Campana, &

Malkewitz, 2010), a fictional brand, SANY, was created and consumers were told that the product was new and would soon enter the market as cover story. Moreover, using an unknown brand would trigger better evaluation of the package among participants since consumers tend to rely more on the package itself when no brand knowledge is present (Magnier &

Schoormans, 2017). The eight stimuli can be found in Appendix 3.

Table 1

Overview of the experimental conditions

Environmental claim No environmental claim Paper/Carton Matte Condition 1:

Paper – Matte – Claim

Condition 2:

Paper – Matte – No claim Glossy Condition 3:

Paper – Glossy – Claim

Condition 4:

Paper – Glossy – No claim Plastic Matte Condition 5:

Plastic – Matte – Claim

Condition 6:

Plastic – Matte – No claim Glossy Condition 7:

Plastic – Glossy - Claim

Condition 8:

Plastic – Glossy – No claim

4.3.1.1 Pre-test

A pre-test was performed through an online questionnaire to examine whether the manipulations within the stimuli were perceived as intended. Before distributing the pre-test, a small sample (N=3) completed the pre-test and acknowledged that the pre-test was clear, and no errors were present. The final pre-test was distributed among the researcher’s own network via informal channels. Eventually, 44 of the 54 participants completed the pre-test and were included in the analysis. This number of participants is higher than the recommended default

39 of 30 participants (Perneger, Courvoisier, Hudelson, & Gayet-Ageron, 2014), and is therefore considered reliable. A within-subjects design was used where participants were randomly assigned to four out of eight stimuli, each followed by four questions: one to test the perceived packaging material, one for surface varnishing and two for environmental claim. To check which packaging material participants considered the package to be, the following question was asked “What material do you think the package was made of?” with a single choice response format, where participants could choose from multiple materials. This question was also used as manipulation check by Petersen and Brockhaus (2017). Furthermore, as taken from Ye et al. (2019), to assess whether the package was indeed perceived as matte or glossy as intended, participants were asked “How would you evaluate the appearance of the package?”

on a 7-point bipolar slider scale (1= matte, 7 = glossy). Lastly, to appraise whether participants noticed the environmental claim but also actually read and remembered the information of the claim, participants were asked to indicate whether they noticed anything else on the package.

They could choose a single option from the following three options: “Yes, a claim stating the material was recycled”, “Yes a claim stating the product is paraben free”, and “No, I did not notice anything else”. Furthermore, for more robustness that the content of the claim was well perceived, a second question was added for environmental claim. Participants were asked to what extent they would agree with the statement: “This package is made of recycled materials”

on a 7-point Likert scale (1 = strongly disagree, 7 = strongly agree). The stimuli used can be found in Appendix 3.

The data was analysed in the statistical software SPSS. To enable analysis, dummy variables were created for all three independent variables. A Chi-Square test was performed for both Packaging Material and the first item of Environmental claim since they were measured on a categorical scale. However, since both tests showed that more than 20 percent of the cells had an expected count less than 5, Fisher’s Exact Tests were performed. The results

40 demonstrated that regarding Packaging Material, the Fisher’s Exact Test statistic of 153.33 resulted in a significant difference between paper and plastic (p=.00). 94.4 percent correctly identified the paper packaging as ‘paper-based/carton’, and 91.8 percent correctly perceived the plastic package as ‘plastic’. Regarding the first item of Environmental Claim, the Fisher’s Exact Test value was 145.50 which represented a significant difference (p=.00) between when a claim was absent, 94.3 percent correctly responded “No, I did not notice anything else”, and when a claim was present, 88.6 percent answered “Yes, a claim stating the material was recycled”. Furthermore, to examine Surface Varnishing and the second item of Environmental Claim, which were numerical variables, Independent Samples T-Tests were performed, for which the homogeneity of variances and normality assumptions were met. Results showed a significant difference in perceived glossiness between the matte conditions, that were perceived as matte (M=2.60, SD=1.48), and the glossy conditions as glossy (M=5.20, SD=1.48); (t(174)=

-11.65, p=.00), with a mean difference of 2.60 (95% CI, p<.05). Concerning the second item of Environmental claim, a significant mean difference of -2.85 (95% CI, p<.05) was detected between when a claim was absent (M=3.14, SD=1.53), where on average participants somewhat disagreed with the package being made of recycled materials, compared to when a claim was present (M=5.99, SD=1.47), where participants on average agreed; (t(174)= -12.62, p=.00). More detailed results of the pre-test are enclosed in Appendix 4.

After the pre-test, 2 participants were questioned, to be certain that the stimuli and measurement items can be used in the actual experiment. This revealed that the stimuli were clear, yet the second item examining Environmental Claim was perceived as a bit confusing.

Therefore, this question is not used in the main experiment, since the other question measured environmental claim well, and therefore the second question is redundant.

41