CONCLUSIONES Y RECOMENDACIONES
4.3. Glosario de términos
Construct validity refers to the validity of inferences about the degree to which the operational measures used (in a survey, for example) accurately reflect the higher order constructs they are intended to measure (Shadish et al., 2002, p. 38). Since it is never possible to establish a one-to-one relationship between the operations of a study and the corresponding constructs, there is always a risk that the degree to which the operations match the constructs will be so poor that the inferences drawn from the data will be invalid (ibid, p. 68). To help reduce this risk, many of the measures used in the survey instrument designed in this study were either exact replications or revised versions of measures used in previous studies by leading researchers in their fields. While there is no guarantee that these previous
measures are perfect, the fact that they were utilized in publications published in leading journals is assumed to be a testament to their (high) quality.
The primary threats11 to construct validity in this study may include the following:
• Inadequate Explication of Constructs: Failure to adequately explicate a construct may lead to incorrect inferences about the relationship between operation and construct.
This threat is particularly difficult to avoid because it is not possible to know how “adequately” a construct has been explicated. In most cases, the constructs utilized in this study have been used in multiple prior studies, and it is believed that the constructs have been more fully explicated over time with each additional study in which they were used. As a result, it is reasonable to be confident in the explications of the constructs and their
legitimacy. However, the research variables in this study (i.e. planners’ commitments and role orientations) are relatively vague concepts and are relatively difficult to quantify and measure. Nevertheless, it is believed that the measures used for these variables are the best available at this time.
The dependent variables used in this study relating to flood hazard mitigation features warrant discussion, for at least three primary reasons. First, this study measures only a subset of all site design techniques that can be used to reduce flood-related losses. This study intentionally focuses only on techniques whose incorporation into development projects can reasonably be expected to be influenced by land use planners. Other flood-loss reduction techniques, such as raising building elevations and strengthening building structures, for example, are more likely to be influenced by other government employees, such as building code officials or engineers. Thus, while the use of such techniques may be important for reducing flood-related losses and may be influenced by the local government as a whole, the
use of such techniques is not likely to be influenced by land use planners, and is thus not examined in this study.
Second, this study measures whether or not particular mitigation site design
techniques were used. It does not control for whether or not the techniques were applicable. Thus, it is not possible to distinguish in this study between techniques that were not used because a decision was made (explicitly or otherwise) not to use them, and techniques that were not used because they were not applicable. For example, projects that did not contain wetlands could not possibly indicate that they preserved wetland buffers. Thus, a project that contained wetlands but did not preserve wetland buffers and a project that did not contain wetlands to begin with would each indicate that they did not preserve wetland buffers. This potential confounding process is a limitation in this study.
Third, certain dependent variables are used to measure whether particular land uses are located completely outside or at least partially inside the floodplain. These variables do not measure the scale of development located inside the floodplain, only the general use. As a result, it is not possible (for example) to distinguish between open space portions of
residential/commercial land uses inside the floodplain and homes/commercial buildings inside the floodplain. The variables used in this study are thus useful for measuring the association between planner characteristics and locating land uses completely outside the floodplain, but are less useful for measuring the association between planner characteristics and the scale of development located inside the floodplain.
While this shortcoming cannot be completely overcome in this study, effort has been taken to provide some verification regarding the nature of land uses located at least partially inside the floodplain. In order to verify that projects with residential and/or commercial land
uses classified as “inside the floodplain” actually contain structures inside the floodplain and not just open space, 11 projects were randomly selected for a preliminary exposure analysis. Google Earth photos, project site plans, and National Flood Insurance Rate Maps (FIRMs) were used to visually verify that structures were located inside the floodplain. Results of this exposure analysis suggest that survey data collected regarding the location of
residential/commercial land uses inside the floodplain are accurate, as Google Earth photos appear to show structures inside the floodplain for each of the 11 projects.
To demonstrate how this analysis was conducted, photos and the FIRM for one of the 11 projects (i.e. The Moorings at Lantana, in Lantana, Florida) are shown here. The
Moorings at Lantana is a redevelopment project located on roughly 10.5 acres that contains 378 dwelling units comprised of condominiums and townhouses. Figure 4.1 shows an aerial photo of the project site, with a rectangle drawn around the project boundary. The waterway to the east of the project is the Atlantic Intracoastal Waterway.
Figure 4.2 shows the FIRM for the Town of Lantana, Florida, with an arrow pointing at the project location. The figure shows that roughly two-thirds of the project site is located in Zone A7, a 100-year floodplain zone.
Figure 4.2 Flood Insurance Rate Map for the Town of Lantana, Florida
As can be seen by comparing Figures 4.1 and 4.2, the portion of the project site located inside the 100-year floodplain clearly contains built structures. Figure 4.3 depicts structures containing condominiums located inside the 100-year floodplain, on the eastern edge of the project site.
Figure 4.3 Structures Containing Condominiums Located Inside the 100-Year Floodplain
Figure 4.4 depicts townhouses located along the southwestern boundary of the project site, roughly one-half of which appear to be located inside the 100-year floodplain.
Figure 4.4 Townhouses at The Moorings at Lantana, Located at Least Partially Inside the 100- Year Floodplain
While this procedure has limitations, it nevertheless provides some degree of
verification regarding the location of development inside the floodplain. While this study is primarily concerned with whether land uses are located completely outside the floodplain, there is nevertheless reason to believe that land uses classified as “at least partially inside the floodplain” do in fact contain built structures.
• Mono-operation Bias: Any one operationalization of a construct both underrepresents the construct of interest and measures irrelevant constructs, complicating inference.
The combination of spatial constraints and a desire to minimize respondent burden served to limit the number of operationalizations that were included in the survey for each construct. As a result, it may be difficult to avoid the threat to construct validity that is posed
by mono-operation bias. However, the operationalizations being used have been used in previous studies, and it is believed that they represent “good” measures of the constructs they seek to represent.
• Mono-Method Bias: When all operationalizations use the same method (e.g., self- reporting), that method is part of the construct actually studied.
This study utilizes a multi-method approach, intended in part to limit the threat to construct validity that is posed by mono-method bias. In addition to the paper survey, informant interviews were conducted with a subset of survey respondents in an effort to corroborate and provide additional meaning to the survey findings.