Statistics Don't (and Can't) Lie

[HankD]

OK I jumped to the end of the thread after reading a few posts.

Statistics don't lie but people do.

I took a statistics course in college and it was surprisingly one of the most difficult I had ever taken.

At on point I had taken both integral and differential calculus, analytical geometry and Vector Analysis and needed a higher math course to assure a general engineering math core so I figured Statistics should be a breeze. NOT!

Long story short, professor walked in on the first day of the course made an opening lecture in which he said (well one of the things) that you can make statistics prove almost anything.

e.g. If a study includes averages you MUST know WHAT METHOD of averaging is being calculated : mean, median or mode and you must also know the range. Are you being told these entities in the study?
No. Then you are a candidate for being snookered.

http://www.purplemath.com/modules/meanmode.htm

Also:
There are also several higher math formulas used in statistics for coming to different final results (lets use polling studies) with the same data depending on related variables which are not usually given in any of the Orwellian press releases fed to we the sheeple depending on which side of the aisle is behind the funding (not always available either).

e.g Weight - one variable in demographics is the percentage of the population in terms of preferential weights.
Does the study (lets say a poll) tell you what the percentage of the general population of those being polled is? (for instance) - 60% Republican, 30% Democrat, 5% Libertarian, 5% other?
Also, what is the random selection algorithm within the range?
Not divulged? Then you are probably being deceived.

And many other statistical math manipulations which we are NEVER (well usually never) told.

Don't get me started on GRAPHS!

HankD

 

[Sapper Woody]

OK I jumped to the end of the thread after reading a few posts.

Statistics don't lie but people do.

I took a statistics course in college and it was surprisingly one of the most difficult I had ever taken.

At on point I had taken both integral and differential calculus, analytical geometry and Vector Analysis and needed a higher math course to assure a general engineering math core so I figured Statistics should be a breeze. NOT!

Long story short, professor walked in on the first day of the course made an opening lecture in which he said (well one of the things) that you can make statistics prove almost anything.

e.g. If a study includes averages you MUST know WHAT METHOD of averaging is being calculated : mean, median or mode and you must also know the range. Are you being told these entities in the study?
No. Then you are a candidate for being snookered.

http://www.purplemath.com/modules/meanmode.htm

Also:
There are also several higher math formulas used in statistics for coming to different final results (lets use polling studies) with the same data depending on related variables which are not usually given in any of the Orwellian press releases fed to we the sheeple depending on which side of the aisle is behind the funding (not always available either).

e.g Weight - one variable in demographics is the percentage of the population in terms of preferential weights.
Does the study (lets say a poll) tell you what the percentage of the general population of those being polled is? (for instance) - 60% Republican, 30% Democrat, 5% Libertarian, 5% other?
Also, what is the random selection algorithm within the range?
Not divulged? Then you are probably being deceived.

And many other statistical math manipulations which we are NEVER (well usually never) told.

Don't get me started on GRAPHS!

HankD

All that's well and good, but this isn't random sampling, or any calculations other than a division of a part by the whole. This is not the type of data that can be manipulated in any way other than non-reporting.

Also, (not trying to be snarky here) I wish you would have read the whole thing. While your post has to do with the title of the thread, it really has nothing to do with the substance of the thread.

 

[HankD]

All that's well and good, but this isn't random sampling, or any calculations other than a division of a part by the whole. This is not the type of data that can be manipulated in any way other than non-reporting.

Also, (not trying to be snarky here) I wish you would have read the whole thing. While your post has to do with the title of the thread, it really has nothing to do with the substance of the thread.

Well, I did look at the data - briefly, but I did and went back to look again just now.
While it looks relatively clean, I have a basic distrust of statistics but if you feel that the data is clean and handled with care then that's good enough for me SW.

HankD

 

[Alcott]

And there is so much evidence in addition to statistics to show there is global warming. But the blind cannot see .... and many who can see refuse to do so.

Have you tried taking the plank out yet?
And why do you post statistics, then, if they are lies?

 

[Alcott]

Maybe you need to spend a little bit more time in that Statistics class? The only thing you should be able to conclude from those numbers is that of the set that was MEASURED, folks may have felt a certain way.

No, you can conclude a lot from statistical sampling. And if you want to sample for300,000,000-- the approximate population of the USA-- you need a sample of 1849 to make a conclusion at the 99% confidence level.

You shouldn't be able to CONCLUDE anything at all about the black population at large from a specific measured set who did a certain thing that the population at large HAS NOT done.

No, not about the entire black population; but yes, about the probability of a particular police incident or criminal incident being done by blacks, by whites, by hispanics, and so on.

 

[Alcott]

You can take a set of statistics and prove just about anything you want.[/QUOTE]

Is that right? Then take the number provided in the original post and prove that most police shootings victimize "native Americans."

 

[Salty]

Dear you hear about the race between the USSR and the USA.

The USSR came in second, and the USA came in next to last.

Who won the race?
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
The USA

 

[HankD]

All that's well and good, but this isn't random sampling, or any calculations other than a division of a part by the whole. This is not the type of data that can be manipulated in any way other than non-reporting.

Also, (not trying to be snarky here) I wish you would have read the whole thing. While your post has to do with the title of the thread, it really has nothing to do with the substance of the thread.

RE:Your reported statistics Sapper Woody.

Yes after further checking I can say that your reported statistics are credible.

My apology for casting doubt on your research.

HankD

 

[Salty]

RE:Your reported statistics Sapper Woody.

Yes after further checking I can say that your reported statistics are credible.

My apology for casting doubt on your research.

My research shows that 76.3% of respondents agree with you

 

[Zaac]

No, you can conclude a lot from statistical sampling. And if you want to sample for300,000,000-- the approximate population of the USA-- you need a sample of 1849 to make a conclusion at the 99% confidence level.

We're not talking about a statistical sample. We're talking about the closed set of people who shot police officers.

No, not about the entire black population; but yes, about the probability of a particular police incident or criminal incident being done by blacks, by whites, by hispanics, and so on.

Wrong again just like yall were months ago. That probability is still 50/50. The only thing you can say about the future of what we're talking about based on a closed set is that this is what happened in the past.

 

[HankD]

My research shows that 76.3% of respondents agree with you

Haha - was that mean, median or mode method and what were the range and weight variables?

HankD

 

[Salty]

Haha - was that mean, median or mode method and what were the range and weight variables?

HankD

I molded the median in a mean way- can you avenge that?

 

[Alcott]

We're not talking about a statistical sample. We're talking about the closed set of people who shot police officers.

And about what that says about the likelihood by race of any particular incident chosen at random.

Wrong again just like yall were months ago. That probability is still 50/50.

P = 0.5 of what?

The only thing you can say about the future of what we're talking about based on a closed set is that this is what happened in the past.

And that says a lot. If you really think a study of where particular crimes occurred, by whom based many differing factors, at what times of day, week, month year, and against whom... you're sillier than I thought.

 

[Zaac]

And about what that says about the likelihood by race of any particular incident chosen at random.

That's the point. It says NOTHING about the population at large. It only speaks to the group of white and black folks who decided to shoot police officers.

You can't use that statistic and say that the Black population in general has got it out for police officers.

P = 0.5 of what?

Of what you mentioned. Either it will be done by someone or it will not be. Two options. 50/50.

And that says a lot. If you really think a study of where particular crimes occurred, by whom based many differing factors, at what times of day, week, month year, and against whom... you're sillier than I thought.

And you're dumber than I thought if you think you can deduce what the general public will do based upon what a group not indicative of the public has already done.

 

[carpro]

Obama is an expert at:

Does it daily.

 

[HankD]

I molded the median in a mean way- can you avenge that?

Yes, expand the time range to eternity.

HankD

 

[Alcott]

And you're dumber than I thought if you think you can deduce what the general public will do based upon what a group not indicative of the public has already done.

"General public"..."not indicative of the public?" So what public are they a part of?