Anyone know much about Stats? Need Help

By joining our community you will have the ability to post topics and access other forums reserved for members. Registration is quick, simple and absolutely free. Join our community today by clicking here.

Master Baiter Joined: 05 May 2008 Posts: 104

Hey, i'm freakin out over trying to analyse my results for my experiment. Need a calmer pair of eyes to have a look.

The hypothesis is that people more intuitive AND in a good mood are more superstitious.

Control= no good mood (puzzle) Exp=good mood (man ball face type vids) Intuitive vs Rational measured by questionnaire. Superstition rated same way.

Initially I thought I would divide the I & R into thirds to categorise them. Top third= rational or Intuitive person. And then do a Chi Squared.

Am I chopping away too much data though? Would a One-Way Mixed ANOVA be appropriate/more informative? And how does an ANOVA work with two distinct conditions?

I'm very confused. Hopefully someone out there is more with it than me today.

Posted: Thu May 29, 2008 1:22 pm

With only one group of test subjects that would normally rule out a one-way Anova. It being for 2 or more groups.
http://en.wikipedia.org/wiki/Analysis_of_variance

You also have no middle ground for those not in good or bad mood, just feeling ambivalent.

Posted: Thu May 29, 2008 1:42 pm

Your not the only one confused... was that english? (sorry I'm just a dumbass car guy.) but if its like a questionare type thing I would be more than happy to help. If its on the techenal side I dont think I can be to much help, (unless you need a engine rebuilt)

Master Baiter Joined: 05 May 2008 Posts: 104

Control is ambivalence. Puzzle isn't frustating, just to pass time neutrally.

the internet is my friend...

I've Independent Sample T-tested the means of Group 1(XP) against group 2(Control). No significance.

Now i'll take the top third of the data. (Inspector gadget, can u remember what that's called, something to do with Quartiles...?) And re-test. Think that sounds ok.

What I really want to do is compare the top third and bottom third against each other and in each group.... Any Ideas?

I'm so shit at my Degree

EDIT: Still shit, but lots of T-Tests should do it. Results are non significant, yay!

Posted: Thu May 29, 2008 4:30 pm

The top third wouldn't be a 'quartile',
Sampling tertiaries is more complicated because you are shortening the range of deviance by definition and the resulting spread would, on an x-y axis, only display rotation around a skewed axis and not reflect against the sum of the whole. Subjective testing of personally supplied results creates two human factor elements which cause their own skew thus further diluting or expanding any data.

Master Baiter Joined: 05 May 2008 Posts: 104

Sampling the tertiaries is to create a category from a scale. To measure people's Intuitive reliance and then mark a certain number of people as reliant or unreliant, middle ground removed to create further space and increase chance of sig diff in a relatively small sample.

Posted: Thu May 29, 2008 6:43 pm

Rackham wrote:

... Control= no good mood ...

Err, speak for yourself, please.

One could also say that acting highly intuitively without necessary control could lead to making compounded mistakes leading in turn to a very, very bad mood.

E.g., not looking and stubbing one's toe; making bad investments and going broke; trusting a scammer and losing money; going to the airport without your passport; etc. etc.

An argument stands and falls with it's premises.

The scientific way is to reach a conclusion based on the data, not to look for an a priori result in the data.

Trophy Videos · Posted: Thu May 29, 2008 6:49 pm

My intuition tells me that your statistical results shall indicate that the exact opposite of your hypothesis is true. For example, take me and my lad, Joe. I am very intuitive and in a good mood most of the time, and I am not one bit superstitious at all. Joe, on the other hand, has demonstrated a complete lack of intuition for the past three months on safari, is in a bad mood most of the time, and is the most superstitious person I have ever encountered. The hypothesis sounds crazy to me. Cool

Baiting Guru Joined: 19 Mar 2007 Posts: 2628

My chi is flowing and I am calm

The answer is 42

So whats your standard deviation?
And are you thinking percentiles.

My answer is yellow

Master Baiter Joined: 05 May 2008 Posts: 104

kleindoofy wrote:

An argument stands and falls with it's premises.

The scientific way is to reach a conclusion based on the data, not to look for an a priori result in the data.

Its all non-significant anyway, and based on very recent research. I'm not looking to make waves in the scientific community, just pass my degree. I'll leave real science to real scientists.

Posted: Thu May 29, 2008 11:05 pm

I am happy most of the time! Smile

Master Baiter Joined: 05 May 2008 Posts: 104

@corona Then you are a lucky lucky person. Very Happy

Posted: Fri May 30, 2008 12:15 am

I am! I have good health, a nice little home, wonderful husband, great family, yard with flowers and all of you. jump_4_joy

What else could I ask for???

Back on topic, which I know nothing about. Embarassed

Master Baiter Joined: 05 May 2008 Posts: 104

nearly finished writing it anyway, don't worry about it. Very Happy

Posted: Fri May 30, 2008 5:24 am

Rackham wrote:

Its all non-significant anyway, and based on very recent research. I'm not looking to make waves in the scientific community, just pass my degree. I'll leave real science to real scientists.

I hope your instructor doesn't find this thread! I was told I was supposed to LEARN something not just throw out some BStuff and get a degree. [/rant]

Stats make me get a terrible rash and want to kick something. Shocked

I hope it worked out well for you. Wink

Master Baiter Joined: 05 May 2008 Posts: 104

I sure hope he doesn't too. He'd be one of the few people on the planet to be able to connect anything I say on here to my real identity.

The way you feel about stats is the way I feel about my degree. All I have to do is print now, and then in 5 1/2 hours I will have completed my course. I'll be a relieved little so-and-so

Posted: Fri May 30, 2008 8:36 am

Rackham wrote:

Its all non-significant anyway, and based on very recent research. I'm not looking to make waves in the scientific community, just pass my degree. I'll leave real science to real scientists.

If your findings either fits with previous research OR you've got something that contradicts previous research and you can give a plausible reason for why our findings differ then you'll be fine. If you're doing a Bachelor's then very little explanation is needed to back up your findings and if it's a Masters you'll need a little more but it's still quite possible to wing it and still be just fine.

Stats glossary:

http://www.stats.gla.ac.uk/steps/glossary/presenting_data.html

Does your data have a normal distribution or what kind of distribution are you working with?

If you've got a nice linear distribution you can always throw in a bit of regression to show off. I'm on holiday so brain is in off mode at the moment. Do you have a binary dependent variable (or whatever you call it in your field) or not?

Good luck!

Posted: Fri May 30, 2008 12:04 pm

Did you factor in the Chos thorie?

Posted: Fri May 30, 2008 12:50 pm

^^^^^^^

Posted: Fri May 30, 2008 1:41 pm

Ill just shut up and go to my room. Sad

Posted: Fri May 30, 2008 2:11 pm

I was asking what you meant, I'm not picking on you (this time Wink

). What is Chos thorie?

Posted: Fri May 30, 2008 3:30 pm

Oops sorry Chaos Theory. I was typing and dealing with a customer at the same time.

Posted: Fri May 30, 2008 3:48 pm

@original poster:

[oops, I see it's too late, but I've written this anyway. You'd better pray I'm not marking your paper!!!!]

I don't know this literature at all but am quite familiar with analysing similar experimental designs. My first thought (as someone who has marked a zillion such experimental reports by undergraduates and postgraduates) is that your hypothesis needs to be clarified in a more specific manner (ideally based on some literature) before you start throwing analyses around.

"The hypothesis is that people more intuitive AND in a good mood are more superstitious."

Do you mean that higher scores on the superstition scale should *only* be observed for those who did the "good mood" task AND are high on the "intuitive" scale, and thus that good mood alone, or intuitive personality alone, is not sufficient to produce an increase on the superstition scale?

If so this seems like a classic interaction hypothesis, where you expect to see no effect of the mood task for "rational" participants, but an effect of the task for "intuitive". 2x2 between subjects ANOVA is the standard approach for a design like this, as long as your data are suitable for anova.

however it's not very convincing to run an analysis which supports your hypothesis, unless there is some other reasonable hypothesis which is rejected by the same analysis. "falsification is more informative than support" as I'm sure you've heard in one form or another.

Posted: Fri May 30, 2008 5:01 pm

I'm not superstitious but very happy! Smile

Trophy Videos · Posted: Fri May 30, 2008 6:37 pm

^^Now I'm not sure if my happiness is because I just because I don't give a sh1t.... Confused

I don't care. I'm smiling. Very Happy

Anyone know much about Stats? Need Help	View next topic View previous topic