Data analysis.

Data were analyzed using SAS (version 9.1.3; SAS Institute, Cary, NC). Frequencies and percentage of distributions are presented for individual items. In addition, a hierarchical linear model was used to compare items within subscales using least squares means. To apply the model, positively phrased items were scored with strongly agree (SA) = 5, agree (A) = 4, undecided/neutral (U) = 3, disagree (D) = 2, and strongly disagree (SD) = 1. Negatively phrased items were reverse scored with SA = 5, A = 4, U = 3, D = 2, and SD = 1. We included negative and positive survey items to keep respondents thinking about each item. The use of mixed negative and positive items reduces the possibility of respondents falling into patterned responses (Stewart et al., 2006). The correlation among responses from the same subject was accounted for using an unstructured covariance structure and the ddfm = KR option was used to adjust denominator df and standard errors.

High desert plant, grass lawn, and tree preference scores were obtained by computing subscale means based on the scoring indicated above. Reliabilities for these subscales were assessed using Cronbach's alpha, yielding 0.90 for desert plants, 0.81 for grass lawn, and 0.73 for trees. Paired t tests were used to compare high desert plant, grass lawn, and tree preferences.

Associations between selected categorical variables were assessed using Fisher's exact test. Associations between binary response variables and a quantitative explanatory variable were explored using logistic regression. All computations were based on available data. Due to item nonresponse, fewer than 99 observations were available for some analyses. Significance was defined at P ≤ 0.05.

Perceptions of high desert plants, trees, and grass lawns.

When high desert plant, tree, and grass lawn mean subscale averages were compared with one another, the tree subscale mean was significantly lower (better) than the high desert plant (P < 0.0001) or grass lawn subscale (P < 0.0001). Grass lawn was the least positively regarded and was significantly higher than desert plants (P < 0.0001). The subscale means and standard deviations based on the n = 93 respondents with complete data were 1.77 ± 0.48 for trees, 2.25 ± 0.73 for desert plants, and 2.72 ± 0.72 for grass lawn (summarized from Tables 1–3). Santa Fe's long history of water conservation and drought management could explain the low acceptance of the traditional grass lawn. Cool-season grasses such as kentucky bluegrass (Poa pratensis) are better suited to the high desert area of Santa Fe than warm-season, drought-tolerant grasses such as bermudagrass (Cynodon dactylon). However, Santa Fe County Ordinance No. 2002–13 prohibits the use of kentucky bluegrass in Santa Fe landscapes (Santa Fe County Water Conservation Information, 2002). Martin et al. (2003) found that urban landscapes in Phoenix, AZ, communities with covenants, codes, and restrictions that mandated the landscape activity had fewer trees and less turf and more shrubs and groundcovers than landscapes in communities without those mandates. Therefore, it is hardly surprising that the traditional lawn represented fewer than 3% of the reported landscape types in Santa Fe (Hurd et al., 2006).

Table 1.

Responses of Santa Fe, NM, respondents (n = 93) to survey items asking feelings about trees in a yard. Participants were asked to give their level of response after being presented with the statement, “The most important reasons for landscaping are…” Survey items are arranged in decreasing order of positive attitude based on least squares means.

View Table

Most respondents agreed or strongly agreed that trees make areas more beautiful, provide valued shade, are a desirable part of the landscape, and increase the value of a home (Table 1). About 74% of the respondents disagreed or strongly disagreed with the statement that trees create too much litter (Table 1). Only 9% of respondents agreed or strongly agreed that trees produce too much pollen. Responses to these items are similar to those reported in a survey of Las Cruces, NM, residents (Spinti et al., 2004), aside from one notable exception: a higher percentage of respondents agreed with trees’ negative aspects. In the Las Cruces, NM, survey, about 33% agreed that trees produce too much pollen, and about 32% agreed that they produce too much litter. One major reason for this difference is that Las Cruces may be perceived to have a high percentage of heavy pollen-producing trees such as mulberry (Morus spp.). Although the City of Las Cruces has not planted mulberries since the early 1990s, well-published, homeowner-driven petitions to ban mulberries from new landscapes in Las Cruces and the prevalence of those trees in existing landscapes (Parks and Recreation Advisory Board, 2004) may have raised homeowners’ awareness of the pollen allergenic potential of those trees.

While the majority of respondents indicated that trees are a desirable landscape feature (Table 1), only about 39% of respondents agreed that the grass is desirable for landscaping (Table 2). Although a preponderance of respondents exhibited some level of agreement with statements that grass is familiar, cooler, provides a play area for children, as well as a place to lie outside, and is a contrast to the desert, 89% also agreed that grass requires a lot of maintenance. We included the ‘grass is familiar’ survey item to gauge the respondents’ familiarity with the grass lawn in an urban setting. The item on the maintenance involved with grass was significantly more negatively rated than any other item and contributed to the overall higher mean of the grass subscale. Respondents were divided on whether grass often looks unattractive, with about 44% agreeing with the statement that grass often looks unattractive, and about 37% disagreeing.

Table 2.

Responses of Santa Fe, NM, respondents (n = 93) to survey items asking feelings about grass lawns in a yard. Survey items are arranged in decreasing order of positive attitude based on least squares means.

View Table

About 78% of respondents agreed that high desert plants provide the landscape they desire (Table 3). This can be compared with the nearly 96% that agreed with this statement for trees (Table 1), and contrasted with the only 39% that indicated that grass is desirable for landscaping (Table 2). People favor landscapes with trees (Schroeder and Cannon, 1987; Spinti et al., 2004), thus it is not surprising that there was strong agreement with the statement that trees provide the landscapes they desired. About 82% agreed that high desert plants look attractive (Table 3). For most respondents, availability and expense of desert plants were not negatives. About 77% disagreed with a statement that high desert plants are not available, and 58% disagreed with a statement that they are too expensive (Table 3). However, the expense of high desert plants does present a concern for some people, as nearly 11% of respondents agreed with the statement that they are too expensive. But taken together, respondents data on variety and cost of desert plants indicate that high desert plants provide sufficient variety at a price point that satisfies most survey respondents.

Table 3.

Responses of Santa Fe, NM, respondents (n = 93) to survey items asking feelings about using high desert plants in landscapes. Survey items are arranged in decreasing order of positive based on least squares means. Survey items had the form “High desert plants…”

View Table

About 64% agreed that high desert plants provide the variety they desire and about 65% agreed that they provide enough green. Even so, about 31% of respondents acceded that high desert plants are not their favorite plants (Table 3). This correlates with an earlier study in which Spinti et al. (2004) found that 39% of homeowners in Las Cruces, NM, indicated that desert plants were not their favorite plants. One possible explanation why desert plants did not appeal to some respondents may be that many desert-adapted plants are grayish green, a color that reflects more heat than bright green (Richards et al., 1986). However, people respond most positively to tree canopies if the foliage approximates the bright green of a young, growing tree (Kaufman and Lohr, 2008). About 58% of respondents expressed some level of disagreement and nearly 22% expressed agreement with a statement that desert plants look too much like the high desert (Table 3). Respondents in Las Cruces articulated similar sentiments, with 44% disagreeing and 25% agreeing with this statement (St. Hilaire et al., 2003).

Current landscaping and willingness to change current landscaping.

When asked to describe the landscaping in their front yard, 64% of respondents indicated southwestern U.S. high desert landscaping, 12% indicated traditional landscaping with turfgrass and non-native trees, and 24% indicated other (Table 4). When asked whether they would be willing to use high desert plants to landscape their front yards, 92% responded that they would. Of those reporting that they currently had desert landscaping in their front yard, only 39% reported that they would like to change their landscaping. This was lower than the percentage of respondents with traditional landscaping or with other types of landscaping who did wish to modify their landscaping (73% and 65%, respectively). The differences among these response distributions were significant, (P = 0.0323), indicating that Santa Fe residents have a preference for retaining their desert landscaping and altering their already low percentage of traditional landscape to more desert-adapted landscape.

Table 4.

Responses of Santa Fe, NM, respondents (n = 99) who reported the actual landscape type in their front yard and back yard.

View Table

For the landscape type in their backyard, 46% of respondents indicated high desert landscaping, 21% traditional, and 33% other (Table 4). Eighty-three percent of respondents indicated a willingness to use high desert plants to landscape their backyard. The relatively lower percentage of residents willing to use high desert plants in their backyard is consistent with the findings of Spinti et al. (2004). The majority of participants in that study showed that they would use desert plant to landscape their front yard (80.3%) and backyard (56.3%) (Spinti et al., 2004).

Respondents also were asked to rate how much tree, shrub, and lawn area vegetation they currently had in their yard by using a scale of none, some, and many. They were also asked whether they would like to have less, the same, or more of those same vegetation types. For trees and shrubs, the most frequent response was some (70% and 71%, respectively) followed by many (28% for trees and 20% for shrubs). Previous studies show that residents in desert environments have a clear preference for trees because of their ability to shade the yard and house (Zube and Kennedy, 1990). For lawn area, the most frequent response was none (63%) followed by some (32%). This result is consistent with the observation that the grass lawn is not a popular landscape element in Santa Fe (Hurd et al., 2006). For trees and shrubs, there was no evidence of an association between the amount they had and the amount they would like to have [Fisher's exact test: P = 0.4050 (trees), P = 0.2049 (shrubs)]. Most people reported that they would like to have the same amount (55.7% trees and 59.8% shrubs). The next most common response expressed was a desire for more (41.24% for trees and 38.14% for shrubs). For the grass lawn, however, there was a significant association between the amount currently in their yard and the desired amount (Fisher's exact test: P = 0.0476). Of those reporting no lawn area, 57% indicated they wanted to keep that same amount, while 29% indicated they wanted more. The 14% who indicated less lawn area (although they had no lawn area), provides an opportunity for an interesting research question that warrants further inquiry. Of those reporting some lawn area, 43% indicated they wanted less, 17% indicated they wanted more, and 40% indicated they wanted the same amount. Only five people had reported a high amount of lawn area, consistent with the fact that water-intensive lawn area is not widespread in residential sites in Santa Fe.

A logistic regression with willingness to use desert plants to landscape the backyard as the response variable, and time in the southwestern United States as the explanatory variable was fit. The fitted equation was logit = 3.0047 − 0.0344 × years, and reflecting a negative relationship between time in the desert southwestern U.S. and a willingness to use high desert plants. This result supports the findings of Spinti et al. (2004) and Martin et al. (2003) who also found a negative relationship between willingness to use desert plants and time in the southwestern United States. Similarly, renters living in a university housing complex on the Arizona State University campus in Mesa report that longer socialization to the desert environment decreased their preferences for arid landscapes.

Our findings are dissimilar to those of Kennedy and Zube (1991) and Lyons (1983) who reported that greater acceptance of desert landscaping is associated with greater familiarity with the desert environment. Those findings contrast with the item “My attitudes about high desert plants have become more positive over time” (Table 5) where nearly 85% of respondents agreed that their attitudes about high desert plants have become more positive over time. Because the current study is cross-sectional, observed associations may be caused by secular trends and demographics. Consequently, older respondents who grew up with most landscaped areas falling into the traditional category may still have a bias toward traditional landscaping, even if they spent much of their life in the southwestern United States. In this survey, we defined traditional landscapes as those containing turfgrass and non-native trees. Respondents who have had shorter residency in the southwestern United States desert may have moved from other areas and become more accustomed to desert plants over time. Furthermore, high desert landscapes have been promoted heavily in recent years due to the ongoing drought in New Mexico, thus, the younger generation might be more influenced by, or more sensitive to, this environmental issue than older generations. Another possible explanation is that younger Santa Fe residents may be more adaptable, and thus more accepting of alternative landscaping. Alternately, the older generation might be sensitive to environmental issues but less able to make changes to their landscapes. Perhaps a longitudinal study of the impact of demographic trends on landscape preferences would parse out this issue.

Table 5.

Responses of Santa Fe, NM, respondents (n = 99) to survey items asking general landscaping attitudes. Survey items are arranged in decreasing order of agreement based on least squares means.

View Table

Water use and factors affecting willingness to reduce water use.

Ninety-three percent of respondents indicated that they water their landscape; the remainder, 7%, indicated they do not. Although the demand for desert-adapted landscapes enjoys a stronger level of support in Santa Fe than in the desert cities of Albuquerque and Las Cruces, it is clear that most Santa Fe residents apply water to their landscapes. What warrants further study is whether residents, mindful of water conservation issues, are watering their desert-adapted landscapes less frequently or are applying less water because of diminished turf areas.

When homeowners who watered their landscapes were asked what factor would cause them to use less water on their landscape, the highest level of agreement was for water shortages. Environmental concerns received the next highest level of agreement (Table 6). Even so, nearly 70% or more agreed with the remaining factors: environmental concerns, city regulations, high water bills, and water rate increases. Overall, the pattern of response to these factors was very similar to that observed by Spinti et al. (2004), although respondents to the current survey indicated stronger agreement to the items on water shortages and environmental concerns. Martin and Kulakowski (1991) and McPherson and Haip (1989) reported that water rate increases were more effective in reducing water use than water shortages. While water rates are critical to long-term water conservation measures (Grisham and Fleming, 1989) and should be used as an urban water-use management tool, our results consistently show that water shortage is the factor most likely to engender a reduction of water use in the landscape. Municipalities should consider weaving the possibility of water shortages into water conservation public awareness campaigns to ensure maximum effectiveness.

Table 6.

Responses of Santa Fe, NM, respondents (n = 99) who use water to survey items asking what would cause them to use less water on their landscape. Survey items are arranged in decreasing order of agreement based on least squares means.

View Table

Landscape preferences.

When asked about landscape preferences, over 88% of respondents agreed that they liked any type of landscape with beautiful healthy plants and that they liked any type of landscape that looks well planned and contains interesting features (Table 5). While the least squares estimate for the item stating “I like any type of landscape that looks neat and well cared for” was higher (representing less agreement) than the least squares estimate for the first two items, 77% of respondents indicated some level of agreement with this statement. About 35% of respondents agreed with the statement “I like yards that provide a contrast to the surrounding desert,” while 31% showed some level of disagreement with this statement. This statement had the least level of agreement among the statements that inquired about general landscape attitudes. These results suggest that plants that blend with the high desert ecosystem and are water-efficient would gain widespread acceptance in managed urban landscapes of Santa Fe. In fact, Nassauer (1988) reported that using plants that fit the ecology of a landscape environment is one way to demonstrate care for the landscape.

In an earlier study, Spinti et al. (2004) showed that residents of Las Cruces, NM, ranked making the yard and house more attractive as one of the top reasons for landscaping. Residents of Santa Fe had the same top reasons when asked about the most important reasons to landscape their yard (Table 7). These reasons are followed in priority by providing a place to relax, increasing property value, and providing shade. A third set of items includes screening the street or neighbors’ yards, expressing one's landscape preferences, and providing a place to play. Based on the least squares mean results, Santa Fe residents had the least agreement with the statement that creating an area of contrast to the desert was the most important reason to landscape their yard (Table 7). About 32% of respondents reported agreement and nearly 47% reported some level of disagreement to this item. While Spinti et al. (2004) reported a higher percentage of neutral responses to this item, based on least squares mean scores, it was the least agreed upon item in their study as well.

Table 7.

Responses of Santa Fe, NM, respondents (n = 99) to survey items asking most important reasons to landscape their yard. Survey items are arranged in decreasing order of agreement based on least squares means.

View Table

The data in Tables 5 and 7 show that landscaping the yard to create a contrast with the desert was the least agreed upon statement. Furthermore, the item in Table 3 stating that high desert plants look too much like the high desert had 22% agreement, but 58% disagreement. A similar item in Spinti et al. (2004) also had greater disagreement (44%) than agreement (34%). Taken together, these data suggest that having managed landscapes contrast with the desert is a low priority for most respondents in our study.