Data overview.

The survey data were gathered through a convenience sampling method, using both online and in-person surveys. The in-person survey was conducted in Central Florida in Oct 2017, and participants were recruited through social media announcements and a master e-mail list developed by the researchers of this study. The national online survey was administered through Qualtrics, a professional online survey platform, from Dec 2017 to Jan 2018. As part of a larger research project investigating pollinator-friendly practices within the ornamental horticulture industry, participants in this study were prescreened to determine if they had purchased plants within the past year and whether they lived in a single-family residence with a yard or garden. Only participants living in single-family homes with landscapes and plant purchase experience were included in the final sample, resulting in 1806 participants. Of these, 126 participants were participants who completed the in-person survey and lived in the local Florida area (Orlando metropolitan area) and 1680 were national online participants (Table 1). The gender ratio was relatively balanced in the online sample (42% male), but it skewed toward a higher percentage of females in the in-person (Floridian) sample, with only 25% male participants. The in-person sample participants had a slightly higher average age of 57 years than the online sample participants (53 years). Both the in-person and online samples consisted of more female and older individuals. This demographic alignment was relevant because individuals 45 years and older are identified as the core consumer segments for lawn and garden products according to the National Gardening Association (2019). Specifically, the 55 to 64 years age group not only reported the highest spending on lawn and garden products in 2018 but also demonstrated the largest growth in expenditures. The in-person sample had a larger household size, more education, and a higher mean household income.

Table 1. Summary statistics of participants’ sociodemographic characteristics and knowledge of native and pollinator-friendly plants.

View Fullscreen

The primary plant purchase outlet for the in-person sample was box stores (71%), whereas only 54% of the national sample reported purchasing from box stores, and 32% purchased from independent garden centers. On average, people in both samples spent between $200 to $300 annually on plants. In addition to sociodemographic information, participants were primarily asked about their plant purchase behavior, knowledge and perceptions of native and pollinator-attractive plants, and knowledge and perceptions of neonicotinoids and relevant regulations.

Subjective and objective knowledge of native and pollinator-friendly plants.

The key variable of interest in this study was the level of consumer knowledge of native and pollinator-friendly plants. Previous studies have shown that perceptions and knowledge of plants, particularly pollinator-friendly plants, can influence purchase decisions and the willingness to pay (Campbell et al. 2017; Getter et al. 2016; Rihn and Khachatryan 2016). Campbell et al. (2017) found that perceived benefits such as attracting pollinators, planting a variety of plant types, as well as liking the appearance of pollinator plants were the major motivations for consumers purchasing pollinator-friendly plants. Despite 30% of the respondents confusing bee-friendly insect management practices with plants that are pollinator-attractive, Getter et al. (2016) showed American consumers were willing to pay a price premium for plants labeled as “grown using bee-friendly insect management practices” or “grown using best insect management practices to protect pollinators’’ compared to other eco-friendly production practices. Additionally, Rihn and Khachatryan (2016) noted that only 24% of surveyed plant purchasers were aware of neonicotinoid insecticides, but those who were aware of neonicotinoids were more knowledgeable about plants that improve pollinator health and also more likely to purchase neonic-free plants. Recently, Wei et al. (2020) measured subjective knowledge of neonicotinoids and both subjective and objective knowledge of pollinator-friendly plants using self-reported rating scores and quiz questions, respectively. They investigated how subjective and objective knowledge interacted with neonicotinoid insecticide-related messages and affected consumers’ willingness to pay for plants grown without neonicotinoids. Similarly, Rihn et al. (2021) investigated how consumers’ subjective and objective knowledge of genetically modified food affected their willingness to pay for genetically modified labeling on food-producing plants.

Similar to Rihn et al. (2021) and Wei et al. (2020), we used both subjective and objective measures of participant knowledge of native and pollinator-friendly plants. To measure subjective knowledge, participants were first asked, “How knowledgeable are you about native or pollinator attractive plants?” based on a 7-point Likert rating scale, with 1 indicating extremely unknowledgeable and 7 indicating extremely knowledgeable. Then, participants were asked to indicate their agreement regarding whether a native variety makes a plant more attractive to pollinator insects using a 7-point Likert rating scale, with 1 indicating extremely disagree and 7 indicating extremely agree. For each quiz question, participants were given two plants with names and images and asked to select the one that was a pollinator attractor. The quiz comprised four questions, and an objective measure was created based on the number of correct answers.

The summary statistics of the knowledge variables are presented in Table 1. The self-rated subjective knowledge scales regarding native and pollinator-attractive plants are consistent. A pairwise correlation of 0.83 suggested that participants who reported being knowledgeable of native plants were likely to report being knowledgeable of pollinator-attractive plants. This pattern was observed for both local and national samples. Only half of the Floridian (local sample) participants agreed that native plants were more pollinator-attractive. In contrast, more than 72% of the national participants believed that native plants were more pollinator-attractive. This difference could be driven by the varying levels of awareness, knowledge, and education regarding native plants and their benefits to pollinators between the local Floridian population and the broader national audience. Florida has a very unique ecosystem and biodiversity. Floridian participants may perceive native plants differently because of their direct experiences with local flora and fauna in central Florida.

Regarding quiz-based objective knowledge, less than 5% of the participants could correctly identify the pollinator-attractive plants presented by the four quiz questions. The pairwise correlations between the objective knowledge variable and three subject knowledge variables were as low as 0.08, 0.05, and 0.10, respectively, revealing that individuals who consider themselves to be knowledgeable (i.e., have high subjective knowledge) tend to overestimate their level of knowledge compared with the results of the objective tests of their knowledge. This result is consistent with existing evidence gap between subjective and objective knowledge (Fernbach et al. 2019; Packard and Wooten 2013; Rihn et al. 2021).

Factors influencing plant purchasing decisions.

To measure the influence of relevant factors on plant purchase decisions, participants were asked to answer the question, “When you purchase plants, please indicate how important the following factors are in influencing your purchasing decisions on plants.” Participants were provided with a list of 15 specific factors, including plant price, appearance, plant care, and sustainability characteristics. These factors were randomly presented to participants (refer to Table 2 for the full list). Participants indicated how important each factor influenced their plant purchasing decisions using a Likert scale of 1 to 7 (1 = very unimportant; 7 = very important). Based on the principal component analysis (PCA), we divided the 15 factors into two broad groups (Table 2). The PCA describes the strength and direction of correlated variables in terms of their potential to quantify unobservable constructs (Jolliffe 1986). The emerging values show the interdependencies between observed independent variables, which can be explained as the indirectly measured, collapsed construct, or component. Details of the loading factors (with loadings >0.6) and Cronbach’s alpha for each of the 15 items can be found in Table 2. “Load” is the term used in the PCA to indicate the highest mean value for each question among all the mean values for the item when testing for linear combinations (Costello and Osborne 2005; Hair et al. 1998). Component values with a Cronbach’s alpha level ≥0.7 are considered to have a strong measure of internal consistency or validity (Tavakol and Dennick 2011).

Table 2. Summary of importance ratings on plant physical and sustainability attributes.

View Fullscreen

Factors such as price, plant size, plant shape, flower/leaf color, plant health, suitability for a garden, and ease of care were categorized as plant physical and care characteristics. Factors such as native plants, pollinator-attractive plants, production methods including using organic practices, using integrated pest management strategies, using non-neonicotinoids pesticides, using no pesticides, using commercial honeybees, and using wild pollinators were categorized as sustainability features. Among the seven items related to plant physical and care characteristics, plant health and suitability for the garden were regarded as essential characteristics across both local and national samples. Rating scores based on the importance of plant sustainability features were generally lower than those of plant physical and care characteristics, especially for (in-person) participants from the local area. Even though both samples reported pollinator-friendly plants as the most important sustainability feature, the mean rating score for the in-person sample was 4.76, which was slightly lower than the mean (5.5) for the national sample.

Empirical framework.

Although the ordered logit model is commonly used to account for the ordinal nature of the outcome variables, it has the restrictive proportional odds assumption or parallel lines assumption, which assume that the effects of explanatory variables on the outcome variable are constant across all levels of the outcome variable (Greene 2003; Train 2003). In other words, the slopes of the regression lines for each level of the dependent variable are equal. However, many empirical studies have demonstrated that the proportional odds and parallel lines assumptions are often violated (Clogg and Shihadeh 1994; Fu 1998; Long and Freese 2006; Peterson and Harrell 1990; Suh et al. 2016). For example, let Y be an ordinal dependent variable taking on M ordered categories, denoted by 1, 2, …, M. Let X be the vector of explanatory variables. According to Williams (2006) and Williams (2016), a generalized ordered logit model can be expressed as a linear function of the explanatory variables:$P\left(Y_i>j\right) = g\left(X{\beta }_j\right) = \frac{\text{exp}\left({\alpha }_j + X_i{\beta }_j\right)}{1+\left\{\text{exp}\left({\alpha }_j + X_i{\beta }_j\right)\right\}}, j = 1,2,\dots ,M-1$

where M is the number of categories of the ordinal dependent variable. From this formulation, the cutoff values of α and the slope parameter β can vary across the value of $j$.

Therefore, the probability that $Y_i$ takes on each of the values are equal to the following:$P\left(Y_i>1\right) = 1\hspace{0.17em}-\hspace{0.17em}g\left(X_i{\beta }_1\right)$$P\left(Y_i = j\right) = g\left(X_i{\beta }_{j-1}\right)\hspace{0.17em}-\hspace{0.17em}g\left(X_i{\beta }_j\right), j = 1,2,\dots ,M-1$$P\left(Y_i = M\right) = g\left(X_i{\beta }_{M-1}\right).$

To determine whether the generalized ordered logit model is a preferred alternative to the ordered logit model, we conducted Brant’s Wald test for the parallel lines assumptions (Brant 1990). As expected, the parallel lines assumption was rejected with a χ² test statistic of 37.32 (P = 0.001) for the factors related to plant physical and care characteristics and a χ² test statistic of 27.52 (P = 0.016) for factors related to plant sustainability features, suggesting that the generalized ordered logit is a better fit for our data structure.

Factors influencing the prioritization of physical and care characteristics versus sustainability features.

The dependent variable of factors related to plant physical and care characteristics was a three-level categorical variable generated based on averaged importance ratings across the seven individual factors within the category. After averaging, the ratings for factors related to plant physical and care characteristics were continuously between 1 and 7. To be consistent with the original Likert scale of 1 to 7, we chose to use 3 as the cutoff point for the unimportant group and 5 as the cutoff point for the important group. Therefore, participant ratings for factors related to plant physical and care characteristics were broadly redefined as unimportant if the average rating across the seven factors was between 1 and 3 (≤3), redefined as neither unimportant nor important if the averaged rating score was more than 3 but less than 5, and redefined as important if the averaged rating score was 5 or more (≥5). Similarly, the dependent variable of factors related to plant sustainability features was a three-level categorical variable generated based on averaged importance ratings across the eight individual factors within the category.

As shown in Table 3, participants in general care about the physical appearance and aesthetic characteristics of plants regardless of whether they are purchasing plants from box stores or garden centers. Nonetheless, participants who primarily buy plants from box stores care less about plant sustainability attributes. In contrast, plant lovers prioritize sustainability attributes. These participants are plant lovers because they are frequent purchasers who spend relatively more money on plants compared to infrequent purchasers. As indicated by marginal effects in Table 4, both plant purchasers who buy from box stores and garden centers are 6 percentage points more likely to rate plant physical and care characteristics as important factors when purchasing plants. Conversely, purchasers who buy from box stores are 9 percentage points less likely to rate sustainability features as an important influencing factor (Table 5). Interestingly, older participants tend to care more about the physical appearance and aesthetic characteristics of plants (Table 3). Male purchasers seem to assign less importance to plant physical and care characteristics and sustainability features compared to their female counterparts. Male purchasers are 4 percentage points less likely to rate physical and care characteristics of plants as important decision factors (Table 4), and they are 5 percentage points less likely to rate plant sustainability as an important decision factor (Table 5). Although there is no difference between online and in-person samples regarding plant physical and care characteristics, online survey participants placed more value on sustainability attributes, was consistent with the higher mean rating scores observed in Table 2.

Moreover, participants who reported higher subjective knowledge place more emphasis on plant sustainability features. Specifically, participants who are more knowledgeable about native and pollinator-attractive plants and report a stronger agreement with the notion that native varieties are more pollinator-friendly prioritized plant sustainability features. Specifically, they were 5 percentage points, 3 percentage points, and 7 percentage points more likely, respectively, to rate plant sustainability features as important factors influencing their plant purchase decisions. Participants’ objective knowledge about pollinator-friendly plants does not impact their prioritization of physical, care, and sustainability characteristics, suggesting a knowledge gap between perceived knowledge and real knowledge. This observation aligns well with and provides support for the findings of Narem et al. (2018), who reported that although subjective knowledge of the environmental benefits of native grass were likely to increase the likelihood of purchase, the objective knowledge (assessed through true/false questions) had little impact on the purchase likelihood.

Further exploration of each factor.

We further explored how participants’ sociodemographic characteristics and knowledge levels may influence how they prioritize each type of physical and care characteristics (Supplemental Table 1.1) and sustainability features (Supplemental Table 1.2) during plant purchase decisions. As shown in Supplemental Table 1.1, plant purchasers who buy from box stores care about both price and plant visual appearance (plant size, shape). In contrast, plant purchasers who buy from garden centers are more likely to consider not only the plant visual appearance but also the plant health, ease of care, and suitability for garden as important factors. Although plant lovers consider physical characteristics as important factors influencing their purchase decisions, they consistently value all the plant sustainability features, including whether plants are native or more pollinator-attractive, whether plants are produced using sustainable production methods (e.g., organic, pesticide-free), or whether plants are produced using pollinators. Additionally, plant purchasers with higher educational levels consider price an important decision factor, but they do not necessarily prioritize individual sustainability features. Regarding individual knowledge of native and pollinator-attractive plants, we only found that individual participants with more subjective knowledge of native and pollinator-attractive plants consistently value all the sustainability features as important factors during their plant purchasing decision-making, but not objective knowledge. These findings are generally consistent with those of existing studies (Narem et al. 2018; Rihn et al. 2023) that have highlighted the importance of perceived knowledge to shaping environmentally conscious behaviors regarding purchasing native and pollinator-friendly plants. Our results also offer additional insights regarding how consumers prioritize physical and nonaesthetic (e.g., ease of care) characteristics versus sustainability features.