Research Design How to do a research project!

Research Design

How to do a research project!

Research designs can be very simple:

Or…. quitecomplicated!

Or… something in between!

A good design willnot compensate forbad fundamentals!

Research Design

A formal written set ofspecifications

and procedures for

There are many ways to set up aResearch Design

So we will look at ageneric plan.

Research Design

A formal written set ofspecifications

and procedures forConducting

and Controlling

a business research project

DESIGN controls for:

Time

DESIGN controls for:

Time

Money

DESIGN controls for:

Time

Money

People

A good DESIGN ensures:

1.The study will be relevant2.That is will use economic procedures

Problem:

The Law of the Instrument

Types of Designs

1.Exploratory Research to gain insight and ideas…

It is for understanding… not for analysis

Types of Designs

1.Exploratory Flexible… good for:a.Diagnosing a situationb.Screening alternativesc. Increase research’s familiarity with problemd.Discovery of new ideas

Types of Designs

Be careful:

Gas lights:

Types of Designs

Be careful:

IBM study in 1947:

“On a humorous note, the principal designer of the Mark I, Howard Aiken of Harvard, estimated in 1947 that six electronic digital computers would be sufficient to satisfy the computing needs of the entire United States. IBM had commissioned this study to determine whether it should bother developing this new invention into one of its standard products (up until then computers were one-of-a-kind items built by special arrangement). Aiken's prediction wasn't actually so bad as there were very few institutions (principally, the government and military) that could afford the cost of what was called a computer in 1947.

Types of Designs

1.Exploratory Flexible… good for:a.Diagnosing a situationb.Screening alternativesc. Increase research’s familiarity with problemd.Discovery of new idease.Gathering background info

Types of Designs

1.Exploratory Methods:a.Situational analysis

Type of Designs

1.Exploratory Methods:a.Situational analysisb.Expert Opinion survey

Types of Designs

1.Exploratory Methods:a.Situational analysisb.Expert opinion surveyc. Literature searchd.Pilot studye.Focus groups

Types of Designs

1.Exploratory2.Descriptive

a. To describe characteristics of a sample

Type of Designs

1.Exploratory2.Descriptive

a. To describe characteristics of a sample b. To estimate proportions

Type of Designs

1.Exploratory2.Descriptive

Cross-Sectional

Types of Designs

1.Exploratory2.Descriptive

Longitudinal

http://www.people-press.org/2013/01/31/trust-in-government-interactive/

Types of Designs

1.Exploratory2.Descriptive

Longitudinal Omnibus

Types of Designs

1.Exploratory2.Descriptive

Longitudinal Panels

Types of Designs

1.Exploratory2.Descriptive

Longitudinal Panels

Problems:Lack of representationThey become experts

Type of Designs

1.Exploratory2.Descriptive3.Associational

What goes with what?

S&P 500

Types of Designs

1.Exploratory2.Descriptive3.Associational 4.Casual (Experiments)

Types of Designs

Casual (Experiments)

Concept of Causality

• Concomitant variation• Time order (casual order)• Elimination of alternative explanations

Types of Designs

Casual (Experiments)

Key is:

Control

Types of Designs

Casual (Experiments)Terms:

Types of Designs

Casual (Experiments)Terms:

Variable

Types of Designs

Casual (Experiments)Terms:

Constant

Types of Designs

Casual (Experiments)Terms:

IV

Independent Variable: treatment exogenous

Types of Designs

Casual (Experiments)Terms:

DV

Dependent Variable: measurement endogenous

Types of Designs

Casual (Experiments)Terms:

Secondary Variables

Something to be controlled thatcould cause the DV to change… Extraneous

Types of Designs

Casual (Experiments)Terms:

Secondary Variables

• Eliminate them• Make them constant• Turn them into IVs• Randomization• Statistical control

Types of Designs

Casual (Experiments)Terms:Field Experiment

Split-plot

Types of Designs

Casual (Experiments)Terms:Laboratory Experiment

Validity Issues

Validity Issues

Internal Validity

The effect is due to IVand

Not to other variables

Validity Issues

External Validity

The effect can be generalized

To the real world

Validity IssuesInternal Validity1. History (Retroactive)2. Proactive History3. Maturation4. Testing Effects

Pre-testing… Post-testing… Interactive effects Reactive measures

Validity IssuesInternal Validity1. History (Retroactive)2. Proactive History3. Maturation4. Testing Effects5. Experimental Mortality

Validity IssuesInternal Validity1. History (Retroactive)2. Proactive History3. Maturation4. Testing Effects5. Experimental Mortality6. Bias

Selection… Interpretation… Etc.

Validity IssuesInternal Validity1. History (Retroactive)2. Proactive History3. Maturation4. Testing Effects5. Experimental Mortality6. Bias7. Statistical Regression

A highly unlikely event is… highly unlikely!

Genetics

Is this statistical regression… or something else?

Validity IssuesInternal Validity1. History (Retroactive)2. Proactive History3. Maturation4. Testing Effects5. Experimental Mortality6. Bias7. Statistical Regression8. Instrumentation9. Luck

Validity IssuesExternal Validity1. Hawthorne Effect

http://en.wikipedia.org/wiki/Hawthorne_effect

Validity IssuesExternal Validity1. Hawthorne Effect 2. Demand Effects

http://allpsych.com/researchmethods/experimentalvalidity.html

Validity IssuesExternal Validity1. Hawthorne Effect 2. Demand Effects

Validity IssuesExternal Validity1. Hawthorne Effect 2. Demand Effects3. Selection Bias

Why?

Why? Rasmussen used “likely voters