Ch. 12: The Correlational Research Strategy Flashcards
the correlational research strategy
involves measuring two or more variables to obtain a set of scores (usually two) for each individual. The measurements are then examined to identify any patterns of relationship that exist between the variables and to measure the strength of the relationship
goal of the correlational research strategy
to establish that a relationship exists between variables and to describe the nature of that relationship
data used for the correlational research strategy
Data usually consist of two or more measurements (one for each of the variables being examined)
individual
refers to a single source, not necessarily a single person
correlational vs. experimental research
- Does not involve manipulating, controlling, or interfering with variables
- This would be experimental research
differences between correlational and differential research
- Correlational research views the data as two scores for each individual and looks for patterns within the pairs of scores to determine whether a relationship exists, while differential design establishes the existence of a relationship by demonstrating a difference between groups
- Correlational studies focus on the relationship between the two variables while differential studies focus on the difference between the groups
labelling pairs of scores in the correlational design
When the data consist of numerical values, the scores in each pair are traditionally identified as X and Y
scatter plot
a graph that presents the two scores for each individual
how do scatter plots represent data?
Represents each individual with a single point
correlation coefficient
a numerical value that measures and describes the relationship between two variables
what does the correlation coefficient describe
the direction, form, and consistency of the relationship
positive relationship
the two variables change in the same direction
negative relationship
the two variables change in opposite directions
linear relationship
the data points tend to cluster around a straight line
pearson correlation
used to describe and measure linear relationships when both variables are numerical scores from interval or ratio scales
monotonic relationship
a relationship that is consistently one-directional, either consistently positive or consistently negative
spearman correlation
used to measure and describe monotonic relationships when both variables are ranks from an ordinal score or have been transformed into ranks
interpreting the consistency of a relationship
- Typically measured by the numerical value obtained for a correlation coefficient
- A coefficient of +1 or -1 indicates a perfectly consistent relationship and a value of 0 indicates no relationship
evaluating relationships for one numerical and one non-numerical score
- Use the non-numerical variable to organize the scores in separate groups. Then, compare the groups using an independent measures t-test (for two groups) or ANOVA (for more than two groups)
- If the non-numerical variable consists of exactly two categories, you can calculate a correlation called a point-biserial correlation
evaluating relationships for two non-numerical scores
- Evaluate the relationship by organizing the data in a matrix, with the categories of one variable forming the rows and the categories of the second variable forming the columns. Then, use a chi-square hypothesis test
- If non-numerical variables consist of two categories, they can be numerically coded as 0 and 1 and then a Pearson correlation can be computer for the data
phi-coefficient
a Pearson correlation computed from coded data
interpreting the phi coefficient
Its numerical value is meaningful but its sign and the concept of a linear relationship are not
how can you evaluate the strength of a relationship?
by computing the coefficient of determination
coefficient of determination
the squared value of a correlation that measures the percentage of variability in one variable that is determined by its relationship with the other variable
how is the coefficient of determination represented?
by r²
guidelines for interpreting the coefficient of determination
small: r= 0.1 or r²= 0.01
medium: r= 0.3 or r²= 0.09
large: r= 0.5 or r²= 0.25
significant relationship
a correlation found in the sample data is very unlikely to have been produced by random variation
sample size and significance
As the sample size increases, it becomes increasingly more likely that the sample correlation accurately represents the real relationship that exists in the population
3 applications of the correlational strategy
- prediction
- reliability and validity
- evaluating theories
prediction and the correlational strategy
- In a correlational study used for prediction, one variable is identified as the predictor variable and the other as the criterion variable (the value being explained or predicted)
- The predictor variable is usually the x value while the criterion variable is the y value
regression
the statistical process of using one variable to predict another
reliability and validity and the correlational strategy
both are commonly defined by relationships that are established using the correlational research design
how is test-retest reliability determined?
by the relationship between an original set of measurements and a follow-up set of measurements
how is concurrent validity determined?
by demonstrating that the scores from the test are strongly related to scores from established tests
strengths of the correlational research design
- describes relationships between variables
- nonintrusive (studies natural behaviours)
- high external validity
- allows researchers to investigate variables that would be impossible or unethical to manipulate
weakenesses of the correlational research strategy
- cannot assess causality
- third-variable problem
- directionality problem
- low internal validity
common use of the correlational research strategy
- Often used for preliminary work in an area that has not received a lot of research attention
- It can describe relationships between variables that might suggest further investigation using the experimental strategy
two limitations in explanations from results of a correlational study
- the third-variable problem
- the directionality problem
the third-variable problem
it is possible that a third (unidentified) variable is controlling the two variables and is responsible for producing the observed relation
the directionality problem
correlational studies do not determine which variable is the cause and which variable is the effect
multiple regression
a statistical technique for studying multivariable relationships
functions of multiple regression
- Allows researchers to examine the relationship between two variables while controlling the influence of the other, potentially confounding variables
- By adding predictor variables one at a time into the regression analysis, it is possible to see how each new variable adds to the prediction after the influence of the earlier predictors has already been considered