# BIOL 1010 Lecture Notes - Internal Validity, Random Assignment, Statistical Significance

55 views3 pages

When the sign of r is negative, variables are said to be negatively related as scores on one

variable increase those for the other tend to decrease. Ex: # of hours spent watching t.v is

negatively correlated with grade point average

Scatter diagrams of positive and negative correlations

In perfect relationships all points fall on a straight line

The values tend to scatter increasingly and become dispersed as the correlation becomes lower

When the correlation reaches 0.00 knowledge of a persons score on one variable tells us

nothing about his or her score on the other

Statistical significance

The magnitude of a correlation coefficient tells us the strength of a relationship between two

variables

Statistical significance refers to the likelihood that the results of an investigation are due to

chance

A statistical significance correlation is one that is not likely to have occurred by chance

A correlation is considered statistically significant if the likelihood or probability that it is a

chance finding is 5 or less in 100. this level of significance is called the .05 level, commonly

written as p =0.5 (p stands for probability)

In general as the size of the correlation coefficient increases the result is more and more likely

to be statistically significant

Whether a correlation attains statistical significance depends on the # of observations made.

The greater the # of observations the smaller r (the correlation) needs to be to reach statistical

significance

Applications to psychopathology

Whenever we compare ppl given one diagnosis with those given another or with ppl without a

psychological diagnosis the study is correlational

When the correlational method is used in research on psychopathology one of the variables is

typically diagnosis

To calculate a correlation between this variable and another one diagnosis is quantifies so that

having an anxiety disorder is designated by a score of 1 and not having one is score of 2

The diagnosis variable can then be correlated with another variable such as the amount of stress

that has been recently experiences

Variables such as having anxiety disorder or not are called classificatory variables

other ex are age, sex, social class and body build. These variables are naturally occurring

patterns and are not manipulated

most research on the causes of psychopathology is corrleational

problems of causality

correlatioanl method has a drawback: it does not allow determination of cause-effect

relationships

a sizeable correlation between 2 variable tells us only that they are related or tend to co vary

with each other but we do not really know which is cause and which is effect or if either variable

is actually the cause of the other

The directionability problem ---ex: a correlation has been found between the diagnosis of shizto

and social class: lower class ppl are more frequently diagnosed as having schitzo

It may be that the disorganized beh patterns of ppl with shitzo cause them to perform poorly on

their educational and occupational endeavours and thus to become impoverished.