**stats**

# Mean accuracy of existing is unlikely due to chance

I seem to get a lot of hits to this post from people googling the phrase **“unlikely due to chance”**. In fact the majority of my traffic comes from this! (SEO worked well, even if it was by accident!) Before this post was probably the most un-useful thing to find if your googling this phrase, as it only consisted of myself moaning about assignments. Seeing as I am a geek for stats I shall explain the phrase so I can stop annoying people.

Basically, when doing statistics, everything is down to probability and chance. The probability of a coin flip is 50/50. There are two events that can happen “heads” and “tails” and both have an equal chance of happening. However, in most statistical analysis it will never be this simple, but what you will want to know is:

Has variable A had an effect on variable B?

OR

How likely is it that the effects observed are due to chance? Could these results have just randomly been picked from a hat of possible results, so in fact there is no affect between variable A and variable B?

The point of conducting controlled experiments and then analysing the data is that you need to prove what you have observed is actually what you have observed. There is no point saying one number is bigger than the other and this proves something, as this could be just due to random chance. You need to prove your results.

A simple example:

*Variable A (amount of sleep) and variable B (test score). Does the amount of sleep you have the night before a test improve the test scores?*

A psychology student is conducting an experiment to see if the amount of sleep you have the night before a test improves the test score. There are two groups, one has their normal amount of sleep and the other group are prevented from sleeping.

On these results you can then perform the appropriate statistical test (in this case an unpaired *t*-test), and find out if the test scores are significantly different depending on the amount of sleep. You will be able to see if they are significantly different from the **p**** value**.

**p**** values** are the probability that a result happened by chance. In psychology for the *p* value to be significant it has to be **p<.05**. If you have a *p* value <.05 this means that there is less than 5% chance of getting the results by chance. Meaning that it is very likely that variable A is, in fact, having an effect on variable B and furthermore very unlikely that the result is due to chance. If the p value is not significant i.e. it is >.05. Your results could have happened by chance and it is likely that there is no effect between the two variables tested.

So…in conclusion **“unlikely due to chance”** is simply when you have tested variables and found a significant* p* value, *p*<.05, the results you have found are unlikely to be due to chance.

ORIGINAL POST – most dull

Finally back to the land of the existing people. I’d say land of the living and be cliche, alas having three large assignments to do is not what I would define as living. My desktop background is nothing more than a taunting ghoul. It consists of all my deadlines, on both my monitors, I’d love to crossed them all off and bask in front of the completed list, or at least pretend for a few minutes. Meh.

Unfortunately, due to my constant germ-age this month I am more behind than I would prefer and will be spending the next three days doing nothing but faux stats, learning theory and some kind of plasticity essay that I have no idea about yet. Not good. Though I have already spent a good 4 hours on my faux stats, it is done bar the fact it needs to be made into English and not just disjointed sentences.

Good news is I finally cracked and bought many digital downloads of David Devant and Mr.Solo, hopefully this will make my assignments much less painful. The joy of a new band I like, well, new to me anyway. Happiness!