Am I missing something here?

[imatfaal]

Just watched a dreadful advert for a men's moisturising lotion - and the claim was that 73% of men agreed it worked. In the small print it said 73% of 50 men in a survey agreed...?

[Prometheus]

Maybe someone only half agreed.

[CharonY]

It is probably one of the 77% of statistics made up on the spot.

[Spellbound]

The 50 men all work for the company, and they just fired the ones who voted no.

[imatfaal]

Maybe someone only half agreed.

 

If I can manage to record the advert so that I can double check then they can explain the wonders of half-agreement to the advertising standards agency. I think you are probably correct - it is a strongly agree, agree, neutral... but it is still bull.

 

It is probably one of the 77% of statistics made up on the spot.

 

I think you have hit the nail on the head - trouble is I have an allergy to lies masquerading as facts.

 

The 50 men all work for the company, and they just fired the ones who voted no.

 

I think you mis-understand - there are always fools who will say yes or no when prompted to; it's the number that annoys me.

 

36 men agree is 72% and one more is 37 men which is 74% - so how do they get 73%?

[studiot]

Was the moisturising lotion London Pride or Fullers Winter?

 

Both these also come in half pints for those who get half cut and make half _ssed statements.

[Phi for All]

What does "worked" mean to the average guy in the context of moisturizing lotion? Whether your skin is dry or normal, a moisturizing lotion should pretty much moisturize the skin. How did it fail to "work" in more than one out of four men? That's some really iffy lotion.

 

We need the survey question. "Did it work, yes or no?" means they completely made up the math and didn't bother to check. "Did it work, yes, no or maybe?" might give you an odd point.

[timo]

Not a native speaker. But maybe a "survey of N" does not necessarily imply that N of the survey group gave a valid feedback. In field tests with customers you can usually be happy if half of the survey group fill out a final evaluation form. In medical trials, there is even explicit statistical methods telling you how to deal with participants that drop out (or drop dead) during the experiment.

[studiot]

 

might give you an odd point.

 

 

73% of 50 is exactly 36.5.

 

Where you count the maybes as yes or no where does the half a response come from?

 

Hence the comments in the other posts.

[John Cuthber]

Some abstained from voting.

[CaptainPanic]

Some abstained from voting.

 

A protest vote on a moisturizing lotion survey?

[tar]

Maybe one was bisexual and they couldn't in fairness to men count the vote as a full man vote, and only gave it half weight as to not be called into question later by someone with imatfaal's need for mathematical accuracy.

Or perhaps they all said it worked and they could only come up with 36.5 fully male amoung them, counting all the halfs. That would be 27 half votes given to the 27 bisexuals in the group and 23 full votes given to the 23 who had their gender figured out as male.

 

After all, they were testing moisturizing lotion.

Something only 75% of the 6 people in my profile picture might have any interest in.

Edited March 25, 2014 by tar

[imatfaal]

Not a native speaker. But maybe a "survey of N" does not necessarily imply that N of the survey group gave a valid feedback. In field tests with customers you can usually be happy if half of the survey group fill out a final evaluation form. In medical trials, there is even explicit statistical methods telling you how to deal with participants that drop out (or drop dead) during the experiment.

 

I think you are almost certainly correct - if they had said x% of n respondents from a survey of N this would have been ok; I think the wording has morphed and no longer fits the data

 

But the terminology was 73% of 50 men surveyed - which is misleading if they have, for example, 25 men saying Yes from 34 respondents taken from a group of 50.

Some abstained from voting.

 

Then the vote should either be shown to be against a postive response to a question ("Do you agree...?) or the voter should not be included in the sample size

 

 

A protest vote on a moisturizing lotion survey?

we gotta fight for the right to have dry skin

 

Maybe one was bisexual and they couldn't in fairness to men count the vote as a full man vote, and only gave it half weight as to not be called into question later by someone with imatfaal's need for mathematical accuracy.

Or perhaps they all said it worked and they could only come up with 36.5 fully male amoung them, counting all the halfs. That would be 27 half votes given to the 27 bisexuals in the group and 23 full votes given to the 23 who had their gender figured out as male.

 

After all, they were testing moisturizing lotion.

Something only 75% of the 6 people in my profile picture might have any interest in.

 

I think we better sweep under the carpet the notion that bisexuals only count as half a man.

[CaptainPanic]

we gotta fight for the right to have dry skin

 

Meh.

But you can get me on the barricades for my conditioner.

[imatfaal]

 

Meh.

But you can get me on the barricades for my conditioner.

 

The next revolution will not be fought under the banner of "liberte egalite fraternite" nor that of "workers of the world unite" - but the chilling clarion call of "Because I'm worth it"

[CaptainPanic]

I have a dream, that one day, your wrinkles will be reduced by 75%!

 

[tar]

imatfaal,

 

Consider it swept.

 

I was only joking around anyway, considering that between my over 80 dad and the over sixty me, we only had enough T between us to count for 3/4 of 2 full male votes. Perhaps if there was a priliminary T count prior the vote, they went by percentage of average male T amoungst the participants to set the weight of a vote, and the actual tally was 73.016% and they rounded down.

 

Regards, TAR2

any way you slice it, they either had a partial man in there somewhere or they did not word their findings correctly

besides I refuse to man the barracades for anything...unless they disparage my right to wear lacey underwear...(did I say that out loud?)

[Prometheus]

I'm sure they just rounded up.

[CaptainPanic]

I'm sure they just rounded up.

 

From 72 to 73?

[imatfaal]

I'm sure they just rounded up.

 

That would be even less legitimate - basically that would be fraudulent advertising; under no circumstance (or under very limited as I am sure someone will think of one even though I cannot) is it cceptable to "round up" exactly 72% (which is 36 men out of 50) to 73%

 

xpost with Capt P

[Prometheus]

Haha, didn't notice. Break time for me. Whatever they did it was dodgy. Do you think you'll ever find out what they actually did?

[imatfaal]

Haha, didn't notice. Break time for me. Whatever they did it was dodgy. Do you think you'll ever find out what they actually did?

 

If I manage to record the advert I will certainly find out - I have a few friends at the Advertising Standards Agency, I also have a preternatural distaste for bunkum masquerading as science; so I will make sure they are asked. Science has a hard enough time convincing "joe public" of the power of empirical evidence without laboratoire garnier and its ilk further muddying the water.