Practical Statistics for Educators - User contributions [en]

Post Hoc Discriminant Analysis

2019-12-07T14:58:53Z

Foxthom4: Created page with "A discriminant function analysis is used to determine which variables discriminate between two or more naturally occurring groups or to "differentiate groups based on quantita..."

A discriminant function analysis is used to determine which variables discriminate between two or more naturally occurring groups or to "differentiate groups based on quantitative variables" (Mayers, Gamst, & Guarino, 2016). Discriminant Analysis can be used to determine what variables might be used to predict the category that a person might belong to.

For example, if a number of observations were recorded about students (e.g., SAT scores, gender, SES, single-parent household) then a research might be able to use a discriminant function analysis to determine whether or not students will fall into one of the following categories: Attends a 4 year college after high school, attends a 2 year college after high school, or does not attend college after high school. The discernment analysis might be used to determine which of the variables are the BEST predictors of a student falling into one of those three groups.

Contributed by: Thomas Fox

Covariates

2019-12-07T14:39:32Z

Foxthom4:

Generally speaking, a covariate is a variable that may impact a dependent variable (other than the previously selected independent variable). A covariate can be controlled for by an analysis of covariance (ANCOVA) or multivariable analysis of covariance (MANCOVA). A real-life example of a study in which a covariate might be included:

Suppose you wanted to know whether or not a certain intervention program had an impact on students' reading ability in grade 2. In this case, the DV is reading ability (as measured by some instrument) and the IV is the intervention. A covariate in this study might be students' reading ability BEFORE placed in the control or treatment group. An ANCOVA could be used to account for this covariate.

''Contributed by: Thomas Fox''

A covariate is used to adjust post-test scores. For example, when completing an experiment, the treatment group may have higher pretest scores than the control group. A covariate would then be used to adjust the post-test scores due to differences in pretest scores. A covariate is one way to control for extraneous variables (Delcourt, 2013).

Delcourt, M. A. (2013) Lecture: Quantitative research designs II. Retrieved from PBworks: http://ed865fall2013.pbworks.com/w/page/68627608/FrontPage.

''contributed by Kara Kunst''

MANCOVA

2019-12-07T14:17:26Z

Foxthom4:

A MANCOVA (Multivariate Analysis of Covariance) is similar to that of a MANOVA (Multivariate Analysis of Variance), but it allows the research to control for covariates (Datallo, 2013). A covariate is an IV, or independent variable, that is not controlled for by the research but could be impacting the DV (dependent variable). Using a MANVOCA can help reduce systematic and in-group error—especially if the sample is not random. An example of when to use a MANCOVA to analysis data follows.

One-Way MANCOVA Question:

''Does the score achieved in the standardized math, reading, and writing test depend on the outcome of the final exam, when we control for the age of the student?
''
DVs: Score achieved on the standardized math, reading, and writing tests
IV: Final exam outcome (pass or fail)

'''Covariate:''' Student age

'''Contributed by: Thomas Fox''' ''Italic text''

MANCOVA

2019-12-07T14:14:15Z

Foxthom4: Created page with "A MANCOVA (Multivariate Analysis of Covariance) is similar to that of a MANOVA (Multivariate Analysis of Variance), but it allows the research to control for covariates (Datal..."

Multiple Linear Regression

2019-10-09T16:06:52Z

Foxthom4: Created page with "Multiple linear regression (multiple regression) is a type of correlational test in which the research is interested in finding the strength of a correlation between multiple..."

Multiple linear regression (multiple regression) is a type of correlational test in which the research is interested in finding the strength of a correlation between multiple variables. In multiple linear regression, multiple variables are used as ''predictors. '''Here, the researcher is interested in the relationship between the predicted variables (dependent) and predictor variables (also known as the independent variables).
''

Independent variables in multiple regression are usually quantitatively measured variables using summative response, interval, or ratio scales (Lawrence, Meyer, & Guarino, 2017)

Multiple Linear Regression uses the same general equation as linear regression, but accommodates for multiple IV's.

Contribution by: Thomas Fox, WCSU Cohort 8

Reference

Lawrence, S., Meyer, G, & Guarino, A.J. (2017). Applied multivariate research: Design and interpretation. Thousand Oaks, CA: Sage Publications

Contributions here

2019-10-09T15:58:26Z

Foxthom4: /* Student Contributors */

== Editor ==
Frank LaBanca, EdD

== Faculty Contributors ==
Karen Burke, EdD

Patricia Cosentino, EdD

Deborah Hardy, EdD

Jennifer Mitchell, EdD

== Student Contributors ==
David Bozzuto

Karen Fildes

Michael Minzloff

Damien Holst

Jennifer Eraca

John Ryan

Kara Kunst

Emily Rhew

Cassandra Cosentino

Kristina Hislop

Mary Fernand

Thomas Fox

Helen Knudsen

Type I and Type II Errors

2019-10-09T15:35:39Z

Foxthom4: Created page with "'''Type One Error:''' An incorrect rejection of the null hypothesis. For example, the researcher falsely states that there is a statistically significant difference between t..."

'''Type One Error:''' An incorrect rejection of the null hypothesis. For example, the researcher falsely states that there is a statistically significant difference between the control group and the experimental group based on their intervention program. This can also apply to correlational tests.

'''Type Two Error:''' An incorrect acceptance of the null hypothesis. For example, the researcher does not report a significance between the control and experimental group based on an intervention when, in fact, there is. In other words, an effect truly exists, but the research reports that there is none.

Contribution by: Tom Fox, WCSU Cohort 8

Lawrence, S., Meyer, G, & Guarino, A.J. (2017). ''Applied multivariate research: Design and interpretation''. Thousand Oaks, CA: Sage Publications