Can statistics be trusted?

[theOperaGhost]

Can statistics really be trusted in research?

I feel they can be at times, however you have to be very careful in exactly what they are saying.

Take this for instance:

There was a study done a few years ago and the conclusion was this: super bowl Sunday is the most dangerous day for women. People could believe this, I suppose...men get all worked up over a football game, drink a lot of beer, etc. But is this study true? No. The true statistics actually said super bowl Sunday is the day domestic violence occurs most. And of course the statement "super bowl Sunday is the most dangerous day for women" came from this because what does everyone associate with domestic violence? Men beating up their wives. Is this the true definition of domestic violence though? No. Domestic violence occurs between/among people who have a special relationship. Domestic violence can be between brothers, father-son, son-father, cousins, brother/brother-in-law, roommates in a living situation, etc, etc, etc...So is super bowl Sunday the most dangerous day for women? It's quite doubtful.

This is just one example of how statistics can be swayed to produce ideal results.

What does everyone else think?

[Modus Operandi]

Ya, they can't always be trusted. I see lots of things on FOX news(go figure:p)that are blatantly misleading statistics.

Most are pretty good, though.

[Sage]

87% of all statistics are made up.

[INFERNO]

There are two main branches of statistics, the branch for interpretative statistics and the branch for seeking patterns in data and such. Both of them are important and necessary, however, the interpretative one is usually the one that is seen as more important and sadly, the one where the outcomes of the actual data can be manipulated.

I do consider statistics necessary for research, although I don't feel that I can honestly say that I would never trust them or that I'd always trust them. Whenever I see statistics in the news, such as FOX, those interpretative statistics I don't trust as much because it's biased people, uneducated in statistics who manipulate the amounts that they understand to fit what they want. Or perhaps this just speaks to the inherit biases from the actual news-reporting groups themselves.

Ideally, if I want statistics, I use my university access to various scientific databases, especially Scopus database to get the actual research studies and read them. I do believe that the researchers can also be biased, however, if I read a bunch of studies, all by different researchers, in different years, who look at the same issue from slightly different aspects, use different methods to obtain the results yet get the same general results, then I'm much more confident.

[theOperaGhost]

There are two main branches of statistics, the branch for interpretative statistics and the branch for seeking patterns in data and such. Both of them are important and necessary, however, the interpretative one is usually the one that is seen as more important and sadly, the one where the outcomes of the actual data can be manipulated.

I do consider statistics necessary for research, although I don't feel that I can honestly say that I would never trust them or that I'd always trust them. Whenever I see statistics in the news, such as FOX, those interpretative statistics I don't trust as much because it's biased people, uneducated in statistics who manipulate the amounts that they understand to fit what they want. Or perhaps this just speaks to the inherit biases from the actual news-reporting groups themselves.

Ideally, if I want statistics, I use my university access to various scientific databases, especially Scopus database to get the actual research studies and read them. I do believe that the researchers can also be biased, however, if I read a bunch of studies, all by different researchers, in different years, who look at the same issue from slightly different aspects, use different methods to obtain the results yet get the same general results, then I'm much more confident.

Agreed (although I absolutely hate that everyone uses fox news as an example...CNN uses biased statistics just as much, but that's a different debate).

Another way research can be misunderstood or misrepresented...

A womens activist group stated that police stations are hiring men over women. When research was done, it was actually proven that the only reason police stations aren't hiring as many women as men is because...guess what! Women aren't applying! Only 11% of the applications for the test are female! Go figure...but no, men are apparently being hired over women.

You've got to watch out for biased researchers and biased statistics...

[ShatteredWings]

87% of all statistics are made up.

Love it


Every single time I do a research or persusasive paper, I bend stats to say whatever the fuck I want them to say.

A fifteen year old writing for school. Is finding facts, and rephrasing them to sometimes say the exact opposite of what was intended to say.

It's not hard.

[INFERNO]

Agreed (although I absolutely hate that everyone uses fox news as an example...CNN uses biased statistics just as much, but that's a different debate).

Well, we could use other examples, although I think we'd then have a pretty long list :lol:.


Another way research can be misunderstood or misrepresented...

A womens activist group stated that police stations are hiring men over women. When research was done, it was actually proven that the only reason police stations aren't hiring as many women as men is because...guess what! Women aren't applying! Only 11% of the applications for the test are female! Go figure...but no, men are apparently being hired over women.

You've got to watch out for biased researchers and biased statistics...

True, researchers can be biased, which is why I find it's best to try to find other studies by other researchers and see if they come up with similar results. If the one researcher gets result A and everyone else gets result B, well, my money is going for result B.

[Atonement]

You can trust statistics, but you can't trust their interpretation.

[RenamedUser]

Statistics, even if entirely accurate, can be selectively chosen, pieced together, and misinterpreted to support practically any theory.

[mrmcdonaldduck]

in one day researching for an asignment i found 3 contridactory stats.

statistics are one of the most reliable sources if found from an unbiased and well researched place but can be extremly unreliable otherwise.