PSYC 2260 Lecture Notes - Lecture 6: January 30, Standard Deviation, Sample Size Determination
26 Jun 2018
School
Department
Course
Professor
PSYC 2260 Introduction to Research Methods in Psychology
Chapter 6 Making Sense of Statistical Significance
Decision Errors
Statistic is about probability, not certainty Thus: always a chance that conclusion is wrong
2 kinds of decision errors:
Type 1 error: rejecting the null when the null is true
- E.g. saying an effect exist when there’s no effect
- P-value is the chance of making a Type 1 error: p=.05 means there is a 5% chance of a Type1
error, called alpha, Greek letter”α”
- Researchers state alpha ahead of time but interpret p-value after data has been analyzed
Type 2 error: not rejecting null when the null is false
- E.g. saying there is no effect when there is an effect, called beta, Greek letter”β”
If we decrease the chance of making a Type 1 error (setting a lower alpha), we increase the
chance
Examples – research implications of Type1 and Type 2 Errors:
Type 1: “dead-end” research
Type 2: premature dismissal of research
Examples – applied implications of Type 1 and Type 2 Errors:
Type 1: needless therapy; waste of time/ resources
Type 2: inaction; withholding and therapy
Consider:
Virus not a threat Virus is a threat
Government get vaccine Type 1
Government says not to worry Type2
- Tradeoffs is risk from Type 1 and Type2 errors can create a dilemma
- Ned a way to minimize both kinds of errors
Effect Size:
Effect size: standardized measure of difference (lack of overlap) between populations.
find more resources at oneclass.com
find more resources at oneclass.com
Document Summary
Statistic is about probability, not certainty thus: always a chance that conclusion is wrong. Type 1 error: rejecting the null when the null is true. E. g. saying an effect exist when there"s no effect. P-value is the chance of making a type 1 error: p=. 05 means there is a 5% chance of a type1 error, called alpha, greek letter . Researchers state alpha ahead of time but interpret p-value after data has been analyzed. Type 2 error: not rejecting null when the null is false. E. g. saying there is no effect when there is an effect, called beta, greek letter . If we decrease the chance of making a type 1 error (setting a lower alpha), we increase the chance. Examples research implications of type1 and type 2 errors: Examples applied implications of type 1 and type 2 errors: Type 1: needless therapy; waste of time/ resources. Tradeoffs is risk from type 1 and type2 errors can create a dilemma.
