Laserfiche WebLink
<br />) <br /> <br />,) <br /> <br />SAMPLING <br />It is recommended that every property be inventoried, at least by a <br />windshield survey, to establish the approximate values and elevations, as <br />well as the appropriate depth. damage relationships to use, Industrial <br /> <br />property and larger commercial businesses require on-site inspections to <br /> <br />determine the value and location of equipment, inventory, and outside <br /> <br />property, For other types of property, sampling is an efficient means to <br />verify windshield estimates of structure and content value as well as <br /> <br />storage locations, <br /> <br />When proper field sampling procedures are used, variances, confidence <br /> <br />interval, and other statistical measures can be derived from the collected <br /> <br />data. This information can assist the planner in developing sensitivity <br /> <br />analyses, in determining optimal solutions, and in evaluating the <br /> <br />confidence that can be placed in study results. If care is not maintained <br /> <br /> <br />in the development and implementation of survey design and sampling <br /> <br /> <br />procedures, biases can be introduced that can lead to spurious results, <br /> <br /> <br />Some common sources of bias that need to be considered when conducting <br /> <br /> <br />sample surveys of floodplain properties include: 1) sampling with unequal <br /> <br /> <br />and unknown stratifications (e.g" only conducting household surveys <br /> <br /> <br />during the daytime, excluding those households from the sample where no <br /> <br /> <br />one is at home during this period); 2) measurement error (e.g" asking for <br /> <br /> <br />only the replacement price paid for a damaged furnace, rather than <br /> <br />obtaining all the information needed to estimate the depreciated <br /> <br /> <br />replacement value of the damaged unit; 3) non. response (e,g., individuals <br /> <br /> <br />who refuse to participate in a survey having different income levels than <br /> <br /> <br />those who do participate; and 4) interviewer error (e,g" a particular <br /> <br />V-25 <br />