Statistics L2

Ethics
 *** not over 1 page
do not over elab

Internal validity
to what extent you can idolate the relationship between the (dv) and the (iv), if you can sucessfully do this you can establihs cause and effect in your experiment. 

Just because a scale is valid doesn't mean it's reliable, this demonstrates that they are two different constructs. 

Published scales are normed as is.

Stats revision

marks feedback which is not an SNG (appres on transcript next to your grades)
Mark for A and B is both 70. 
you can't say you've performed equally well in both tests, you need another piece of information. 
I would liek to know how other people went on average (the mean) 
The mean is a measure of central tendancy. 
Revision of mean, median and mode. 

We need some kind of measure of variability of dispersion. 
What is considered the operation? 
2 would tell you something about the variability of that number set? 

What is central tendancy, if you end up with 0 that can't be right because there's variability between 4 and 6. so the simplist thing you can do is drop the sign (-).

The def of variability is how far on average are the dumbers from eachother? 
Making 1 the new index of variability. 

You've worked out that the mean of the data set is 3 and now you're going to subtract it from all the numbers.
5-3
4-3
3-3
2-3
1-3 
Does this column have a label? 
This column is called the column of devaitions from the mean. So it's called a table of deviation scores. 
Then draw any postivies from the negaitive signs. 
Third column is called squared deviation scores. 
we want an average so we divide by 5 and is 2 a legitimate index of variation for this data set? 
yes and does it have a name? 
2 is the variance of this data set. 
But it suffers from 1 problem it's in squared units. 
It should be in the same units ans what you measured so you square root it. and it's square root is called standard deviation. 

Which test did I perform better in?
score - mean over SD
This result is called the z-score.

population
Is defined by the researcher
let's ***ume that the distribution is normal
if you're plotting height how can you ***ume a normal distribution wouldn't there be outliers or funny shapes or shouldn't the date be represented like a correlational graph with a line through it?

Instead of asking how many cm's taller I am than the population mean I could always ask how many SD's am I above the mean. But why would I want to ask it that way?
which you would find by x- mew over sigma. (Apparently this is also a z score and I fail to understand why) 

We can find the distance to the area under the curve using a z table. 
Why is it better to use a z table for this, why do…