SOKOLOVA, KRISHNA, AND D € ORING
485
TABLE12
STUDY 4B: ATTRIBUTE UTILITIES
Plastic condition
Plastic þ paper condition
M (SD)
t -Stat.
p -Value
M (SD)
t -Stat.
p -Value
(Test value ¼ 0)
(Test value ¼ 0)
Zero-centered utility scores a
Price
68.98 (25.74) 47.33 (83.06) 84.13 (49.50)
38.00 8.08 24.10
< .001 < .001 < .001
66.66 (24.60) 29.51 (86.45) 90.20 (47.55)
38.42 4.84 26.89
< .001 < .001 < .001
Montperal brand b
Weight
a $2.00 price, Lays and Tyrrells brands, and 3 oz weight served as the baseline attribute levels for utility estimates. b On average, Montperal is 47.33 utils worse than the other two brands when it comes in plastic packaging; it is 29.51 utils worse when it comes in plastic þ paper packaging.
ANOVAs to compare attribute utilities across the “plastic- control” and “plastic-sticker” conditions. The analysis revealed that price had a significant nega- tive utility, and the disutility of price was greater in the “plastic-control” condition ( M plastic-control ¼ 62.74, SD ¼ 28.33, vs. M plastic-sticker ¼ 56.87, SD ¼ 27.46, F (1, 588) ¼ 6.54, p ¼ .011, g p 2 ¼ 0.011). Flavor had a signifi- cant positive utility, meaning that chocolate chips were preferred, on average, to nuts. There was no significant dif- ference in the utility of flavor across the “plastic-control” and “plastic-sticker” conditions ( F (1, 588) ¼ 0.57, p ¼ .449, g p 2 ¼ 0.001). Next, we analyzed the effect of the sticker intervention on the utility of packaging. In the “plastic-control” condi- tion, the average utility of plastic packaging was negative and significantly different from zero ( M plastic-control ¼ 10.98, SD ¼ 43.49, t (294) ¼ 4.33, p < .001), suggest- ing that, on average, people were less prone to choose the bar in plastic than the bar in plastic þ paper. Thus, the lower choice propensity for plastic compared to plastic þ paper packaging, observed in the between-subjects studies 4a and 4b, manifested in a within-subjects design as well. Critically, in the “plastic-sticker” condition, the utility of plastic became positive and significantly different from zero ( M plastic-sticker ¼ 35.05; SD ¼ 61.48, t (294) ¼ 9.79, p < .001), suggesting that when the sticker intervention was introduced, on average, people were more prone to select the product in plastic packaging over the product in plastic þ paper packaging. Thus, with the sticker intervention, there was a reversal in preference for plastic over plastic þ paper. Finally, the difference in packaging utilities was signifi- cant across the two intervention conditions ( F (1, 588) ¼ 110.22, p < .001, g p 2 ¼ 0.158), suggesting that the bar in plastic packaging was significantly less likely to be chosen in the “plastic-control” condition compared to the “plastic- sticker” condition. Perceived Environmental Friendliness. Next, we ana- lyzed the PEF scores ( a ¼ 0.95) of plastic packaging across the “plastic-control” and “plastic-sticker” conditions. A
choice propensity for products packaged in plastic com- pared to products packaged in plastic þ paper. Study 5 introduces a managerially relevant intervention that aims to alleviate this bias in environmental friendliness perceptions, making consumers value plastic packaging more than plastic þ paper packaging. In addition, the study tests whether, holding everything else equal, we will also observe the relative advantage of plastic þ paper packag- ing when people choose between plastic- and plastic þ paper-packaged products. Method Students at a public university ( N ¼ 590, M age ¼ 19.93, 51% female) were randomly assigned to one of two condi- tions (intervention: plastic control vs. plastic sticker). The method was similar to that adopted in studies 4a and 4b. Participants made 12 choices between two granola bars. Half the participants saw granola bars packaged in plastic þ paper or in plastic only (plastic-control condition). The remaining participants saw the bars packaged in plastic þ paper or in plastic with a sticker “USDA certified minimal packaging” (plastic-sticker condition; see table 13 for screenshots of sample choice tasks). In addition to packag- ing, the bars were described on price ($3.00 vs. $3.50 vs. $4.00) and flavor (chocolate chips vs. nuts). After making 12 choices between granola bars, partici- pants saw the plastic-packaged granola bar (without or with a sticker) and rated its packaging on the four-item PEF scale. Finally, they reported their age and gender.
Results
Conjoint Analysis. We obtained individual-specific utilities of each attribute—price, packaging, and flavor— using the hierarchical Bayes procedure in Sawtooth sepa- rately for the “plastic-control” and “plastic-sticker” condi- tions (table 14). Packaging and flavor attributes were dummy coded with plastic þ paper packaging and nuts fla- vor serving as the baseline. Price was coded linearly ($3.00 ¼ 0; $3.50 ¼ 1; $4.00 ¼ 2). We then ran one-way
Made with FlippingBook flipbook maker