Upgrade your account to a Community Supporter account and remove most of the site ads.

Reply to thread

Message: <blockquote data-quote="BigFreekinGoblinoid" data-source="post: 979498" data-attributes="member: 5988">As long as every voter clearly understands the difference between a "0" and an "I don't know", then things should be fine. As a market researcher, I have found that a 5 grade ranking scale is the most widely accepted, however. This is a minor point though. Data "cleansing" is another typical part of the market research process. In fact many surveys purposely let you answer  ( the same ) questions in an inconsistant manner so you can discount any one trying to fudge ( or just being careless ) with the survey. Voters should realize that if they answer 10/0/0/0/0 then their vote will NOT be counted, as this is obviously bad data to input into any survey results. 10/2/2/2/2 should be discarded too etc...10/DK/DK/DK/DK would be perfectly acceptable however. Generally you will need some type of Ecxel ( or Access etc... )data tabulation to compile results. Example: results will look like this after tabulation and cleansing:Best Dealy-wig I use all the time: Product A: Total Votes: 940  Total Points: 7050Average Score:  7.5  Product B: Total Votes:  329 Total Points: 2698Average Score:  8.2  Product C:  Total Votes: 551 Total Points: 4353Average Score  7.9 Product D:  Total Votes:  441Total Points: 3837Average Score 8.7 Product E:   Total Votes:  767Total Points: 6136 Average Score  8.0 So Product D, while perhaps only 3rd in the regognition/familiarity ( not "popular" ) vote, wins with a superior average rating. The only problem you have in this type of voting is if you do not get a large enough vote for any one item in any category. Market research guidelines stipulate that you need at least ( about ) 200 inputs for any question to get above an 80% confidence rating ( 450 responses gets you 90% -/+ 5% )-BFG</blockquote>

Verification