Research Methods
for the digitally inclined
by 
Stephen R. Schmidt

 

Basic Methods II

Rival Hypotheses

I.  Burden of Proof

II.  Two Basic Research Designs

III.  Threats to Internal Validity

IV.  Summary

I.  Burden of Proof

A.  Burden of proof:

For any statement in science (psychology) lies on the investigator.  If I say that X causes Y, then it is up to me to convince you of this fact.

B.  Proof requires:

That I not only show how X and Y are related, but that I show that that relation exist with all else equal.
 

C.  Alternative Explanations:

Demonstrating that other factors are not responsible for the observed relation between X and Y requires that I rule out any plausible alternative explanations of the findings.
 
Alternative Explanations = Rival Hypotheses

II.  Two Basic Research Designs

A.  Within Subjects Comparisons


Observation 1 - Treatment - Observation 2

Example:

Homework    -    Reward    -    Homework
 

B.  Between Subjects Comparisons

Group 1 (control):          -           Observation
Group 2 ( Exp.):     Treatment    Observation
 

C.  Each design has its own strengths and weaknesses.


Each design is susceptible to different threats to internal validity.
 

III.  Threats to Validity

A.  The seven mortal sins

 (Catholic Church)
  1. Avarice
  2. Gluttony
  3. Envy
  4. Pride
  5. Anger
  6. Lust
  7. Sloth

B.  Eight mortal sins of a researcher

- each represents an alternative explanation of your results

M  I  S    S  M  I  T  H

  1. Maturation
  2. Instrumentation
  3. Selection
  4. Statistical Regression
  5. Mortality
  6. Imitation of Treatments
  7. Testing
  8. History

Maturation

Any systematic changes that occur over time, and are thus confounded with the treatment.
Hypothesis:

Observe     Treatment   Observe

Alternative:

Observe     time         Observe

Instrumentation (Instrument Decay):

Changes that occur in the measuring device (instrument) during the course of the research that are confounded with treatment.
 
Hypothesis:

Observe             Treatment                 Observe

Alternative:

Observe           Instrument Change         Observe

Selection

Systematic differences between the comparison groups as a result of how participants were placed in (selected for) the groups.
 
        Hypothesis:                           Alternative:
Group 1 :     Control                                   1st    sub-population
Group 2 :     Treatment                               2nd   sub-population
 

Statistical Regression

Between a first and a second assessment, extreme scores tend to shift toward the mean or the average.
 
Hypothesis:

Observe             Treatment                       Observe

Alternative:

Observe             Statistical Regression     Observe

Example:  Give children a test of reading ability.  Select those children who score particularly low on the test (below the mean).  These children are then placed in a reading program.  After the program is over, the reading test is administered again.
How it works


    An individuals scores  on a test varies from test to test.

    If you specifically select participants who scored low on a first test Ö.

    Ö it is most probable that they will score higher on the next test.

Mortality (Attrition)

Loss of research participants during the course of the study contaminate the results.
Hypothesis:


Observe             Treatment                       Observe

Alternative:

Observe               Loss of Participants      Observe

Example:  smoking cessation

Imitation of Treatments (carry-over)

The results of a treatment effect are allowed to contaminates (carry-over) to other conditions.
Hypothesis:                  Alternative:

Treatment 1                  Treatment 1

Treatment 2                  Knowledge of Treatment 1

Testing

Repeated measurement of the same individual may impact a personas score on the test.
Hypothesis:

Observe             Treatment                                Observe

Alternative:

Observe              Impact of first observation     Observe

History

Events that take place during the course of research that influence the outcome of the study.
Hypothesis:

Observation        Treatment              Observation

Alternative:

Observation        (historical event)   Observation

IV.  Summary

A.  The burden of proof is on the researcher to rule out alternative explanation of results.

B.  There are eight major threats to internal validity that represent alternative explanations.

C.  These can be summarized by the acronym:  MIS SMITH

  1. Maturation
  2. Instrumentation
  3. Selection
  4. Statistical Regression
  5. Mortality
  6. Imitation of Treatments
  7. Testing
  8. History
About these pages | Syllabus & Lecture Notes | Projects | Sample Tests | Related Links | Exit