12.01.2015 Views

RESEARCH METHOD COHEN ok

RESEARCH METHOD COHEN ok

RESEARCH METHOD COHEN ok

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

STATISTICAL SIGNIFICANCE 517<br />

hands and the size of feet. The task is not to support<br />

the hypothesis, i.e. the burden of responsibility is<br />

not to support the null hypothesis. If we can<br />

show that the hypothesis is not supported for 95<br />

per cent or 99 per cent or 99.9 per cent of the<br />

population, then we have demonstrated that there<br />

is a statistically significant relationship between<br />

the size of hands and the size of feet at the 0.05,<br />

0.01 and 0.001 levels of significance respectively.<br />

These three levels of significance – the 0.05,<br />

0.01 and 0.001 levels – are the levels at which<br />

statistical significance is frequently taken to have<br />

been demonstrated, usually the first two of these<br />

three levels. The researcher would say that the<br />

null hypothesis (that there is no statistically<br />

significant relationship between the two variables)<br />

has not been supported and that the level of<br />

significance observed (ρ) is at the 0.05, 0.01 or<br />

0.001 level. Note here that we have used the<br />

terms ‘statistically significant’, and not simply<br />

‘significant’; this is important, for we are using<br />

the term in a specialized way.<br />

Let us take a second example. Let us say<br />

that we have devised a scale of 1–8 which<br />

can be used to measure the sizes of hands and<br />

feet. Using the scale we make the following<br />

calculations for eight people, and set out the results<br />

thus (see http://www.routledge.com/textbo<strong>ok</strong>s/<br />

9780415368780 – Chapter 24, file 24.9.ppt):<br />

Hand size<br />

Foot size<br />

Subject A 1 1<br />

Subject B 2 2<br />

Subject C 3 3<br />

Subject D 4 4<br />

Subject E 5 5<br />

Subject F 6 6<br />

Subject G 7 7<br />

Subject H 8 8<br />

We can observe a perfect correlation between the<br />

size of the hands and the size of feet, from the<br />

person who has a size 1 hand and a size 1 foot<br />

to the person who has a size 8 hand and also a<br />

size 8 foot. There is a perfect positive correlation<br />

(as one variable increases, e.g. hand size, so the<br />

other variable – foot size – increases, and as one<br />

variable decreases so does the other). We can<br />

use the mathematical formula for calculating<br />

the Spearman correlation (this is calculated<br />

automatically in SPSS):<br />

r = 1 − 6 ∑ d 2<br />

N(N 2 − 1)<br />

where d = the difference between each pair of<br />

scores, = the sum of, and N = the size of<br />

the population. We calculate that this perfect<br />

correlation yields an index of association – a<br />

coefficient of correlation – which is +1.00.<br />

Suppose that this time we carry out the<br />

investigation on a second group of eight people<br />

and report the following results:<br />

Hand size<br />

Foot size<br />

Subject A 1 8<br />

Subject B 2 7<br />

Subject C 3 6<br />

Subject D 4 5<br />

Subject E 5 4<br />

Subject F 6 3<br />

Subject G 7 2<br />

Subject H 8 1<br />

This time the person with a size 1 hand has a size 8<br />

foot and the person with the size 8 hand has a size<br />

1 foot. There is a perfect negative correlation<br />

(as one variable increases, e.g. hand size, the<br />

other variable – foot size – decreases, and as one<br />

variable decreases, the other increases). Using the<br />

same mathematical formula we calculate that this<br />

perfect negative correlation yields an index of<br />

association – a coefficient of correlation – which<br />

is −1.00.<br />

Now, clearly it is very rare to find a perfect positive<br />

or a perfect negative correlation; the truth<br />

of the matter is that lo<strong>ok</strong>ing for correlations will<br />

yield coefficients of correlation which lie somewhere<br />

between −1.00 and +1.00. How do we<br />

know whether the coefficients of correlation are<br />

statistically significant or not (See http://www.<br />

routledge.com/textbo<strong>ok</strong>s/9780415368780 –<br />

Chapter 24, file 24.10.ppt.)<br />

Let us say that we take a third sample of eight<br />

people and undertake an investigation into their<br />

hand and foot size. We enter the data case by case<br />

(Subject A to Subject H), indicating their rank<br />

Chapter 24

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!