Mary Jackson Pitts, Ph.D.

 

Hit Counter

mpitts@astate.edu

Home ] Up ]

Research ethics

Dr. Mary Pitts

 Ethics

n     Collection of data

u  The measurement tool     

« Post hoc Research Questions

n     Training data collectors

n     Drawing the sample

n     Inputting data

n     Disclosing purpose. Full exposure not always wise.  Can bias results

 

Inputting data

n     Cleaning up data

u  Possible values assigned to responses

 

Analyzing data

n     Setting level of significance at p< .05

u  p.=.055

    Is it significant?

 

Treating respondents  correctly

n     Purpose of IRB

n     Debriefing of respondents

Informed consent

u  Must have title of project

u  Benefits to Participants

u  Description of risk

u  Statement of Confidentiality

u  Participation is voluntary

u  Contact information

Recognizing error

n      Limitations

Measurement error

n     Instrumentation

n     Application

n     Sampling error

n     Random sources of error

Reliability

n     Internal consistency reliability

n     Alternative forms reliability

n     Split half reliability

u  First half last half split

u  Odd even reliability

n     Balanced halves and random halves.

n     Test retest

Validity

n     Internal Validity

n     External Validity

Internal validity

u  Instrument measures what it is suppose to measure.

u  Internal validity questions conclusions

External Validity

n     External validity is concerned with generalizing to other things.

u  Replication

u  Randomization

Internal validity sources

n     History

n     Maturation

n     Testing

n     Instrumentation

n     Statistical regression

n     Subject selection

n     Mortality

n     Selection maturation interaction

Three types of validity

n     Content validity

n     Criterion or predictive validity

n     Construct validity

u  Known group validity

External validity sources

n     Reactive effects of testing

n     Selection X experimental variable interaction

n     Reactive effects of experimental arrangements

n     Multiple treatment interference