- Home
- Find a Therapist
- Topic Streams
- Get Help
Mental Health
Addiction
ADHD
Anxiety
Asperger's
Autism
Bipolar Disorder
Depression
Eating Disorders
Insomnia
OCDPersonality
Passive Aggression
Personality
ShynessPersonal Growth
Happiness
Goal Setting
Positive PsychologyRelationships
Low Sexual Desire
Relationships
SexEmotion Management
Anger
Procrastination
StressFamily Life
Adolescents
Child Development
Elder Care
Parenting
SiblingsRecently Diagnosed?
Diagnosis Dictionary
- Magazine
- Tests
- Psych Basics
- Experts
You've got to accentuate the positive,Eliminate the negative,And latch on to the affirmative.Don't mess with Mister In-Between- Johnny Mercer (1944)
This essay is not about positive psychology, false or true, although I hope the deliberately ambiguous title drew some of you readers here. Read More
Sigh
My hero.
RE: "the credibility of psychology as a scientific field is at stake"
I'm accentuating the positive to elicit your smile this Thanksgiving.
An artistic satire from Tom Lehrer that fits your essay "like a glove!"
Tom Lehrer on Sociology
http://www.youtube.com/watch?v=wX5II-BJ8hI&feature=related
Enjoy! =)
False false positive
"Remember the statistical significance level of p < .05, which means that a given finding could have occurred by chance only one out of twenty times."
Not quite. A statistical significance level of p<0.05 means that a given finding (e.g., a difference of means such as the observed or larger) could have occurred only one out of twenty times supposing the null hypothesis is true . In probability terms, this means that the alpha level is a conditional probability: the probability of observing such data given that the hypothesis is true, or P(Data|Ho=TRUE). That is, when we calculate the p value for a given test, we assume that the null hypothesis is true. The question we make is, "if we suppose that the null hypothesis is true, how likely are we to get a result such as this one or larger?"; the answer is the p value. From this, the classical inference trick is to decide to accept or reject the null if the probability of obtaining such result under the null is small (the famous 0.05 alpha level).
There are plenty of problems with this classical view (many more than the ones well discussed by Simmons et al., and there are many (myself included) that do not think this is a very good inferential process, but at least we ought to get our facts straight. This false interpretation of the p value has been criticised well since the 60s and 70s, but researchers outside of the methodology departments in the social behavioural sciences have mostly managed to remain unaware of it.
Post new comment