Chapter 5 - Alternatives to Experimentation

What is correlational research?

Research in which no independent variables are manipulated and relationships between two or more measured variables are examined.

What can correlational research be used for?

it can be used to predict relationships between variables, but it cannot establish causation.

What is a predictor variable and a criterion variable?

A predictor variable (X) is used to predict the value of a criterion variable (Y)

ex of a predictor variable and criterion variable

You want to see if hours spent studying (X) can predict exam score (Y).

• Predictor variable (X): Hours spent studying

• Criterion variable (Y): Exam score

• If students who study more tend to get higher scores, study time predicts exam performance.

Why can't correlational research establish cause-and-effect relationships?

because there is no manipulation of an independent variable, and problems such as the third-variable and directionality problems can occur

What is the third-variable problem?

A third variable may be responsible for the relationship observed between two variables.

What is the directionality problem?

When two variables are related, it is unclear which variable influences the other.

What are three situations in which correlational research is useful?

During the early stages of research, when manipulating a variable is impossible or unethical, and when studying naturally occurring variables.

ex of a situation in which correlational reserahc is useful

Naturally occurring variables: You want to study whether age and memory ability are related.

What are three factors that can affect a correlation?

1. outliers

2. curvilinear relationships

3. truncated variables

What is the coefficient of determination (R²)?

The proportion of variance in one variable accounted for by another variable.

What R² value is considered a strong correlation according to Cohen (1988)?

R² ≥ .25

What is linear regression used for?

To find the straight line that best fits data plotted on a scatterplot and use it for prediction.

What is the least squares regression line?

The best-fitting straight line through the data.

What is the standard error of estimate?

The difference between the predicted and observed values of Y

What do a, b, and Ŷ represent in a regression equation?

a = y-intercept, b = slope, Ŷ = predicted score

What is a multivariate design?

A design involving multiple dependent variables analyzed together

What is the difference between univariate and multivariate analysis?

Univariate = one dependent variable

multivariate = multiple dependent variables.

What is MANOVA?

Multivariate Analysis of Variance, used to analyze multiple dependent variables in an experiment.

Why can MANOVA be preferable to multiple separate ANOVAs?

It analyzes multiple dependent variables in one analysis and is more powerful than multiple univariate ANOVAs.

What are the major assumptions/requirements of multivariate statistics?

Linearity, absence of outliers, normality, homoscedasticity, absence of multicollinearity, and large sample size.

What is homoscedasticity?

The variability in scores on one variable is roughly the same across values of another variable.

What is heteroscedasticity?

Unequal variability across values of another variable; on a scatterplot it can appear as a conical pattern.

What is multicollinearity?

When two or more dependent variables are highly correlated, often because they measure the same concept.

What is factor analysis used for?

To reduce a large number of variables into smaller sets of related variables

What is a factor and what is a factor loading

A factor is a group of related variables representing a common underlying dimension

a factor loading indicates the strength of a variable's relationship with that factor.

What is the difference between exploratory and confirmatory factor analysis?

Exploratory identifies how variables cluster when there are no predetermined ideas

confirmatory tests whether variables cluster as predicted.

What is the difference between principal components and principal factors analysis?

Principal components reduces variables and summarizes data

principal factors is used when the research is based on theoretical or empirical predictions.

what is the difference between partial and part correlation

Partial correlation removes the effect of a third variable from both variables; part correlation removes it from only one variable.

What is multiple regression?

A form of regression used to examine the relationship between multiple predictors and a dependent variable

What are the three types of multiple regression?

1. Simple: all variables entered at once.

2. Hierarchical: variables entered in an order determined by theory/model

3. Stepwise: order determined statistically.