Inferential statistics (Pearson Product-Moment Correlation Coefficient)

Purpose of inferential statistics

To examine relationships, make predictions, and determine differences among groups.

Considerations/factors that determine appropriateness of techniques

study purpose, hypotheses, research question, lvl of measurement of the variables, design, number of groups studied

What are the four key factors in using an algorithm for inferential statistics?

Nature of the research question or hypothesis, level of measurement of DV, number of groups studied, and research design (independent or paired).

non-parametric study design

- difference or association

- nominal or ordinal level variables

- skewed/severe kurtosis (even with continuous variables)

- independent or paired samples

parametric study design

- difference or association

- interval/ratio level variables

- independent or paired groups

independent sample

subjects are unrelated to other subjects or observations

paired sample

subjects are related in someway to other subjects or observations

regression analysis

prediction of a DV using one or more IV

Pearson-Moment Correlation Coefficient

expression of the relationship or association studies (2 methods: spearman rank, pears product-moment (r))

Pearson correlation coefficient

examines relationship between 2 continuous variables which are measured at the interval or ratio level (NOT determine cause and effect just associations)

direction of relationships

positive or negative

positive direct relationship

1 variables increases as the other increases, or 1 decreases as other decreases

negative indirect relationship

1 variable increases as other decreases

Strength of Relationships

weak, moderate or strong

weak relationship

less than 0.3

moderate

between 0.3 and 0.5

strong

greater than 0.5

r value of 0.00

No relationship or association between the two variables.

degrees of freedom for pearson's r

N-2

Effect Size

Pearson r is EQUAL to the effect size or strength of relationship between 2 variables

strenght of ES

Same values as R values (smaller ES = greater SS needed, larger ES = smaller SS can be used)

role of research questions in descriptive and correlational studies

direct the focus of the study on examining relationships, making predictions, or determining group differences.

negative r value indicates

indirect relationship where one variable increases as the other decreases.

percentage of variance

increase understanding about relationships between 2 variables in terms of clinical importance --> stronger the r value, greater the percentage of variance

calculating variance

half of r value ( r= 0.5, then 25% of variance in one variable is explained by the other) (r^2)x100