Correlation A Research
Correlation A Research
Correlation A Research
By Mike Rippy
studies may be used to A. Show relationships between two variables there by showing a cause and effect relationship B. show predictions of a future event or outcome from a variable
Observational Research e.g. class attendance and grades 2.Survey Research e.g. living together and divorce rate 3. Archival Research e.g.violence and economics
It allows the researcher to analyze the relationship among a large number of variables 2. Correlation coefficients can provide for the degree and direction of relationships
Interpreting Correlations
Scattergram-
a pictorial representation of correlations between two variables Use of a scattergram An x and y axes are produced perpendicular to each other Results of correlates are plotted The relationship of these plots are interpreted
is multiple correlation (0 to 1) (b) is regression weight which is a multiplier added to a predictor variable to maximize predictive value B is beta weight which is used in a multiple regression equation to establish the equation in a standard score form
when comparing two variables sometimes inference may be made that one causes the other. Only an experiment can provide a definitive conclusion of a cause and effect relationship.
tend to break down complex patterns into two simple components. Researcher identify complex components that interest them but could probably be achieved in many different ways.
Prediction Studies
A
variable whose value is being used to predict is known as the predictor variable A variable whose value is being predicted is the criterion variable. The aim of prediction studies is to forecast academic and vocational success.
more variables to the model Replicate design Convert question to the experimental design
extent to which a criterion pattern can be predicted Data for developing a theory for determining criterion patterns Evidence about predicting the validity of a test
large of a sample may cause faulty data to occur 15 to 54 people should be sampled per variable used.
research in useful for practical purposes Definitions- selection ratio- proportion of the available candidates that must be selected Base rate- percentage of candidates who would be selected without a selection process
Correlation- Used when both variables are expressed as continuous scores Correlation Ratio- Used to detect nonlinear relationships
Correlation Coefficient
It measures the magnitude of the relationship between a criterion variable and some combination of predictor variables Correlation coefficient of determination equals R squared. This expresses the amount of variance that can be explained by a predictor variable of a combination of predictor variables
can range from 0.00 to 1.00. The larger R is the better the prediction of the criterion variable. There is more statistical significance if the R squared value is significantly different from zero.
Canonical Correlations
Is
when there is a combination of several predictor variables used to predict a combination of several criterion variables
Path Analysis
Is
a method of measuring the validity of theories about causal relationships between two for more variables that have been studied in a correlational research design
Correlation Matrix
Is
an arrangement of row ad columns that make it easy to see how measured variables in a set correlate with other variables in the set
a method of multivariate analysis that test causal relationships between variables and supplies more reliable and valid measures than path analysis It is also called LISREL which stands for Analysis of Linear Structural Relationships
Differential Analysis
This
is subgroup analysis in relationship studies This application is used when the researcher believes that correlated variables might be influenced by a particular factor. Then subjects from the sample are selected who have this characteristic
are times when a certain test is more valid in predicting a subgroups behavior. The variable that is used in this instance is called a moderator variable