You can make statistics show what ever you want, .

Do it all the time.
I always ask people what kind of answer they want.
Then, if there is data which is “Spoiling it”, I ask them to find a reason why it should be disregarded from the study as erroneous outliers.

“6 out of 10 Cats prefer Dolphin Sick than any other brand”. You know how they did that?
They ask thousands of cat owners. Randomly. They MUST keep the survey random to fall within the advertising rules. But, they search down the whole survey until they find a couple of groups of 10 consecutive answers with 6 voting for Dolphin sick. 1 group of 10 is a chance. 2 is a pattern so it can be used. Despite the fact it is statistically (And sensibly) unsound.