Advertisement: In a carefully controlled study, blindfolded volunteers were divided evenly into five groups.
Conclusion (this shows that)
Sparkle Cola elicits a more favorable response from consumers than any of the competing colas tested.
Evidence
Each person in the study tasted Sparkle Cola and one of five competing colas (each group a different cola).
Most of the volunteers said they preferred Sparkle Cola to the competing cola tasted.
Evaluate
Given that ... most of the volunteers picked Sparkle as their preference over the other cola they tasted,
How can we argue that ... Sparkle does not elicit a more favorable response from consumers than any of the competing colas?
The argument is moving from, , so the author is definitely assuming that the sample in this study was big enough, representative enough, unbiased, etc, that it can be trusted as a sample that reflects consumer preferences.
But beyond the Sampling concerns, did the author successfully establish that Sparkle elicited a more favorable response from the study volunteers than did any competing cola tested?
If Soda X were one of the sodas tested and it was actually the #1 favorable response from these study participants, then how come most participants picked Sparkle as their preference?
Well, most participants didn't even get a chance to try Soda X. Only one of the five groups did. Maybe Soda X kicked butt within its group, but Sparkle kicked butt in the other four groups. That could create a situation where most participants ended up picking Sparkle, even though it looks like Soda X is more preferred.
If numbers help, suppose there are 20 people in each of the five groups, and let's say this is how the voting went down.
Soda V (9) vs. Sparkle (11) Soda W (8) vs. Sparkle (12) Soda X (20) vs. Sparkle (0) Soda Y (5) vs. Sparkle (15) Soda Z (6) vs. Sparkle (14)
52 out of the 100 people (most) said they preferred Sparkle to the competing soda. But if we looked at this table of data, we'd still think that Soda X was the winner. Sparkle beat out the other four sodas, but Soda X demolished Sparkle.
Goal
Look for an answer discussing the Sampling move or pointing out that even within the study, it's possible that most votes went to Sparkle even though one of the sodas crushed Sparkle in its head to head.