Statistical Tables The statistical tables are labeled across the top (i.e., the banner) with the respective cross-tabulation descriptors (bannerpoints such as male, female, Western Washington, Balance of West). Below these Bannerpoint descriptors are the bases (the number of respondents) used to calculate the columns of percentages. Columns of percentages which add to more than 100% are the result of computer-rounding errors or multiple responses. Small differences from 100% are usually computer-rounding errors, while large differences typically are the result of multiple responses. Statistical Variation. All percentages shown in the statistical tables are subject to statistical variation, or statistical error. The smaller the sample of respondents (i.e., the smaller the "base"), the larger is the statistical variation in the corresponding percentages, usually. The table below presents approximate statistical error for percentages based upon different sample sizes. Statistical Error Ranges* Size of Sample 50 75 100 150 200 250 300 400 500 600 700 800 900 1000 50% 14.0 11.4 9.8 8.0 6.9 6.2 5.6 4.9 4.3 4.0 3.7 3.4 3.2 3.0 At Various Percentage Levels 40% 30% 20% or or or 60% 70% 80% 13.7 11.1 9.6 7.8 6.8 6.1 5.5 4.8 4.3 3.9 3.6 3.3 3.1 3.0 12.8 10.4 9.0 7.3 6.3 5.7 5.2 4.5 4.0 3.6 3.3 3.1 2.9 2.8 11.2 9.1 7.9 6.4 5.5 4.9 4.5 3.9 3.5 3.2 2.9 2.7 2.6 2.4 10% or 90% 8.3 6.8 5.9 4.8 4.2 3.7 3.4 2.9 2.6 2.4 2.2 2.0 1.9 1.8 *At 95% level of confidence. When sample sizes are small (less than 100), extra caution should be exercised in interpreting the corresponding percentages. Decision Analyst, Inc. 80