Bruce Nunnally Posted August 28, 2015 Report Share Posted August 28, 2015 I enjoy Chris' tests Bruce 2023 Cadillac CT4-V Blackwing Follow me on: Twitter Instagram Youtube Link to comment Share on other sites More sharing options...
cdgrinci Posted August 29, 2015 Report Share Posted August 29, 2015 I agree with Chris, depending on the cleanliness of the engine before adding the solution, that would have the most effect on results. Looking at the data, and the standard deviation, my conclusion is that results are a wash and within normal distribution (no tangible gain). And as Chris says, results will vary. Chuck '19 CT6, '04 Bravada........but still lusting for that '69 Z-28 Link to comment Share on other sites More sharing options...
Bruce Nunnally Posted August 29, 2015 Author Report Share Posted August 29, 2015 I think your analysis is correct -- I suppose what was missing from the test was a placebo group. The amazing part of the test was that Chris got so many people to record at all!! but to use a placebo group you might send 8 oz of unleaded gas in an unmarked jar to a random group, and this product in the same type jar to another random group. That way you can compare the results from the 8 oz of gas to the results from the additive? Bruce 2023 Cadillac CT4-V Blackwing Follow me on: Twitter Instagram Youtube Link to comment Share on other sites More sharing options...
winterset Posted August 29, 2015 Report Share Posted August 29, 2015 Seems unrealistic that someone's mpg would increase 11mpg. I would have tossed out that data. Also with the margin of error, the results are definately the same. Link to comment Share on other sites More sharing options...
Bruce Nunnally Posted August 29, 2015 Author Report Share Posted August 29, 2015 I do like the idea of "crowd-sourcing" tests but it gives new issues of how to ensure consistency. How would you solve those? Bruce 2023 Cadillac CT4-V Blackwing Follow me on: Twitter Instagram Youtube Link to comment Share on other sites More sharing options...
winterset Posted August 31, 2015 Report Share Posted August 31, 2015 I would recommend using the standard deviation / throwing out values outside 95% confidence. That 11 mpg increase was an obvious error which skewed the results. Link to comment Share on other sites More sharing options...
Bruce Nunnally Posted August 31, 2015 Author Report Share Posted August 31, 2015 I would recommend using the standard deviation / throwing out values outside 95% confidence. That 11 mpg increase was an obvious error which skewed the results. That would be a reasonable quantitative way to prevent outliers from effecting the result. Bruce 2023 Cadillac CT4-V Blackwing Follow me on: Twitter Instagram Youtube Link to comment Share on other sites More sharing options...
Recommended Posts
Archived
This topic is now archived and is closed to further replies.