Laserfiche WebLink
User Guide to The National Citizen SurveyT" <br />Non - response Bias <br />Knowing that residents in single family dwellings are more likely to respond to a mail survey, NRC <br />oversamples residents of multi - family dwellings to ensure their proper representation in the sample <br />data. Rather than giving all residents an equal chance of receiving the survey, this is systematic, <br />stratified sampling, which gives each resident of the community a known chance of receiving the survey <br />(and apartment dwellers, for example, a greater chance than single family home dwellers). <br />Weighting <br />The first step in preparing the data for analysis is to weight the data to reflect the demographic profile <br />of the residents of the community being surveyed. Weighting is the approach used by quality survey <br />consultancies to ensure that the demographic characteristics of the sample mirror the overall <br />population. It is an important method to adjust for potential non - response bias. NRC uses a special <br />software program of mathematical algorithms to calculate the appropriate weights. Several different <br />weighting "schemes" may be tested to ensure the best fit for the data. <br />" "Don't know" Responses <br />Generally, a small portion of respondents select "don't know" for most survey items and inevitably some <br />items have a larger "don't know" percentage. Comparing responses to a set of items on the same scale <br />can be misleading when the "don't know" responses have been included. If two items have disparate <br />"don't know" percentages (2% versus 17 %, for example), any apparent similarities or differences across <br />the remaining response options may disappear once the "don't know" responses are removed. Such an <br />example is shown below. <br />When comparing the community as a place to live to the community as a place to work, it would appear <br />that 76% of respondents rated the community as a place to live as "excellent" or "good" compared to <br />just 63% for the community as a place to work. However, the community as a place to work has a much <br />higher proportion of respondents answering "don't know" (17% compared to 2 %). <br />Place to live Place to work <br />Number Percent Number Percent <br />Excellent 48 25% 38 20% <br />Good 97 51% 81 43% <br />Fair 23 12% 22 12% <br />Poor 19 10% 17 9% <br />Don't know 3 2% 32 17% <br />Total 190 10 0% 190 10 0% <br />If we remove the three "don't know" responses from the community as a place to live and the 32 "don't <br />know" responses from the community as a place to work, the two items are actually much more similar <br />in their evaluations: 78% "excellent" or "good" place to live compared to 75% "excellent" or "good" place <br />to work. <br />Place to live <br />Place to work <br />Number �— <br />Percent <br />Numbe <br />Percent <br />Excellent <br />48 <br />26% <br />38 <br />24% <br />Good <br />97 <br />52% <br />81 <br />51% <br />Fair <br />23 <br />12% <br />22 <br />14% <br />Poor <br />19 _ <br />10% <br />17 <br />11% <br />Total <br />187 <br />100% <br />158 <br />100% <br />