Order #403985 Topic: #3

tutorthammy
Chapter_5_and_6.pptx

An Applied Reference Guide to Research Designs Quantitative, Qualitative, and Mixed Methods

A Visual System for Research Designs in

Education and the Social & Behavioral Sciences

© 2012 SAGE Publications

All Rights Reserved

Edmonds, W. A., & Kennedy , T. D. (2012). An applied reference guide to research designs: Quantitative, qualitative, and mixed methods. Thousand Oaks, CA: Sage.

Ch. 5 - 6

Factorial Designs

Solomon n-group designs

Sage Publications

Factorial designs

Can be either between- or within-subjects

Or can include a combination both between- and within-subjects. These are called mixed-subjects

Extension of the k-factor design

Includes a minimum of 2 factors and each with at least 2 levels

The simplest form of a factorial is the 2 x 2 or the 22 (sk).

Multiply the number of factors and levels to determine the number of groups or cells required:

2 x 2 = 4 groups; 2 x 3 x 2 = 12 groups (or cells)

© 2012 SAGE Publications

All Rights Reserved

Family of factorial designs

Factorial designs can have mixed levels

2 x 3 (two factors- one with two levels and one with 3 levels)

These designs can have more than two factors

3 x 3 x 3 (3 total factors each with 3 levels)

Can contain both pre and posttests, but most just contain a posttest

A 2 x 4 design can utilize a within-subjects or a between-subjects approach depending on the research scenario

A 2 x 3 design can be a mixed-factorial design

The first factor at 2 levels is tested as between-subjects and the second factor at 3 levels is tested as within-subjects

© 2012 SAGE Publications

All Rights Reserved

Different types of factors (i.e., independent variables)

An independent variable from a design perspective is the variable that can be manipulated

Hence the IV is controlled

An intervention can be controlled via inclusion and exclusion for example

Gender in a factorial design is an IV, but cannot be “manipulated” (thus, sometimes referred to as a quasi-IV) by the researcher the same way an intervention is manipulated

However, gender can be controlled via statistical procedures and portioned into male and female

© 2012 SAGE Publications

All Rights Reserved

Experimental Research for Factorial designs

These designs can be designated as experimental research (cause-effect) so long the following can be ensured:

Covariation?

Temporal Precedence?

No Plausible Alternative Explanation?

Threats to internal validity can vary depending on:

The approach (between-, within-, or mixed-subjects)

The sequencing of observations (posttest only or if a pretest is included)

© 2012 SAGE Publications

All Rights Reserved

Most common threats

Most common threats to internal validity related, but not limited, to these designs:

Experimental. Maturation, Testing, Diffusion, and Instrumentation

Quasi-experimental. Maturation, Testing, Instrumentation, Diffusion, and Selection Bias

If a factorial design includes a within-subjects component then Sequencing Effects becomes a threat to internal validity

© 2012 SAGE Publications

All Rights Reserved

2 x 2 factorial Design Between-subjects

© 2012 SAGE Publications

All Rights Reserved

Research Question Example Here: Fill in the blanks (consider the main and interaction effects)

Formulate a RQ which would necessitate the use of this type of design

Group Treatment Posttest
1 XA1B1 O1
2 XA1B2 O1
3 XA2B1 O1
4 XA2B2 O1
Time ►
Assignment Group Factor (A) Factor (B) Posttest
1 (A1) (B1)
2 (A1) (B2)
3 (A2) (B1)
4 (A2) (B2)
Time ►

2 x 2 factorial Design Between-subjects

What are the general weaknesses (in terms of internal validity) related to this design?

What are the strengths (in terms of internal validity) of this design?

Is it advisable to utilize random assignment with this design? Why or Why not?

How does one decide how to manipulate the factors?

© 2012 SAGE Publications

All Rights Reserved

2 x 2 factorial Design within-subjects

© 2012 SAGE Publications

All Rights Reserved

Research Question Example Here: Fill in the blanks (consider the main and interaction effects)

Group or Individual Treatment Posttest
1 XA1B1 O1
1 XA1B2 O1
1 XA2B1 O1
1 XA2B2 O1
Time ►
Group Factor (A) Factor (B) Posttest
1 (A1) (B1)
1 (A1) (B2)
1 (A2) (B1)
1 (A2) (B2)
Time ►

2 x 2 factorial Design within-subjects

What are the general weaknesses (in terms of internal validity) related to this design?

What are the strengths (in terms of internal validity) of this design because of the removal of the pretest?

How would sequencing effects be controlled for in this design?

How does one decide this should be conducted as a within-subjects approach?

© 2012 SAGE Publications

All Rights Reserved

2 x 2 x 2 factorial Design between-subjects

© 2012 SAGE Publications

All Rights Reserved

Research Question Example Here: Fill in the blanks (consider the main and interaction effects)

Group Treatment Posttest Group Treatment Posttest
1 XA1B1C1 O1 5 XA1B1C2 O1
2 XA1B2C1 O1 6 XA1B2C2 O1
3 XA2B1C1 O1 7 XA2B1C2 O1
4 XA2B2C1 O1 8 XA2B2C2 O1
Time ► Time ►
Assignment Group Factor (A) Factor (B) Factor (C) Posttest
1 (A1) (B1) (C1)
2 (A1) (B2) (C1)
3 (A2) (B1) (C1)
4 (A2) (B2) (C1)
5 (A1) (B1) (C2)
6 (A1) (B2) (C2)
7 (A2) (B1) (C2)
8 (A2) (B2) (C2)
Time ►

2 x 2 x 2 factorial Design between-subjects

What are the general weaknesses (in terms of internal validity) related to this design?

What are the strengths (in terms of internal validity) of this design because of the removal of the pretest?

How does one decide how many factors to use and how many levels each factor is split into?

How many cells are required for a 2 x 2 x 2 design?

© 2012 SAGE Publications

All Rights Reserved

2 x 2 x 2 factorial Design mixed-subjects

© 2012 SAGE Publications

All Rights Reserved

Research Question Example Here: Fill in the blanks (consider the main and interaction effects)

Group Treatment Posttest Group Treatment Posttest
1 XA1B1 C1C2 O1 5 XA1B1 C2C1 O1
2 XA1B2 C1C2 O1 6 XA1B2 C2C1 O1
3 XA2B1 C1C2 O1 7 XA2B1 C2C1 O1
4 XA2B2 C1C2 O1 8 XA2B2 C2C1 O1
Time ► Time ►
Assign Group Factor (BS) (A) Factor (BS) (B) Factor (WS) (C) Posttest
1 (A1) (B1) (C1) and (C2) Test
2 (A1) (B2) (C1) and (C2) "
3 (A2) (B1) (C1) and (C2) "
4 (A2) (B2) (C1) and (C2) "
5 (A1) (B1) (C2) and (C1) "
6 (A1) (B2) (C2) and (C1) "
7 (A2) (B1) (C2) and (C1) "
8 (A2) (B2) (C2) and (C1) "
Time ►

Factor C is treated as the WS variable

What are the general weaknesses (in terms of internal validity) related to this design?

What are the strengths (in terms of internal validity) of this design because of the removal of the pretest?

Is it advisable to utilize random assignment with this design? Why or Why not?

How does one decide which factor is the between-subjects and which factor is the within-subjects?

© 2012 SAGE Publications

All Rights Reserved

2 x 2 x 3 Factorial design

Mixed-subjects

Solomon N-group Design

A between-subjects approach

An extension of the factorial design

Combines pretest only and the pre- and posttest design

Controls for a major threat to internal validity and construct validity

Testing

Pretest sensitization

Unique research scenarios warrant the use of the Solomon N-group Design

How can group equivalency be tested?

Should there be a certain amount of time that should elapse before the next groups receives the intervention? Why or why not?

© 2012 SAGE Publications

All Rights Reserved

Solomon N-group Design

Solomon 4-group Design

1 factor

Solomon 6-group Design

2 factors

Solomon 8-group Design

3 factors

© 2012 SAGE Publications

All Rights Reserved

Solomon N-group Design

Most common threats to internal validity related, but not limited, to these designs:

Experimental. This design controls for all threats to internal validity except for Instrumentation.

Quasi-experimental. Instrumentation and Selection Bias

© 2012 SAGE Publications

All Rights Reserved

Solomon four-group design

© 2012 SAGE Publications

All Rights Reserved

Research Question Example Here: Fill in the blanks

Group Pretest Treatment Posttest
1 O1 XA O2
2 O3 - O4
3 - XA O5
4 - - O6
Time ►
Assignment Group Pretest Treatment Posttest
R 1
R 2 -
R 3 -
R 4 - -
Time ►

What are the general weaknesses (in terms of internal validity) related to this design?

What are the strengths (in terms of internal validity) of this design because of the removal of the pretest?

Is it advisable to utilize random assignment with this design? Why or Why not?

What are the logistical concerns that would keep a researcher from utilizing this design?

© 2012 SAGE Publications

All Rights Reserved

Solomon four-group design

Solomon six-group design

© 2012 SAGE Publications

All Rights Reserved

Research Question Example Here: Fill in the blanks

Group Pretest Treatment Posttest
1 O1 XA O2
2 O3 XB O4
3 - XA O5
4 - XB O6
5 O7 - O8
6 - - O9
Time ►
Assignment Group Pretest Treatment Posttest
1
2
3 -
4 -
5 -
6 - -
Time ►

What are the general weaknesses (in terms of internal validity) related to this design?

What are the strengths (in terms of internal validity) of this design because of the removal of the pretest?

Is it advisable to utilize random assignment with this design? Why or Why not?

What are the logistical concerns that would keep a researcher from utilizing this design?

© 2012 SAGE Publications

All Rights Reserved

Solomon six-group design

Access companion website to further explore specific references of interest

http://www.sagepub.com/edmonds/

Be cautious of internet resources on research design and methodology.

Be aware that many books use different types of terminology to explain the same thing.

© 2012 SAGE Publications

All Rights Reserved

Group 1

O1

X

O2

Group 2

O1

-

O2

Group 1

O1

X

O2

Group 2

O1

-

O2