SELECTING A HYPOTHESIS TEST: a decision chart
Based on Robert S. Witte (1993) Statistics 4th Ed. New York: Harcourt Brace
RETURN
edited 8/19/11
DATA TYPE |
Condition 1 |
yes/no |
Condition 2 (if needed) |
yes/no |
Testing Process |
|
Condition3: Are assumptions violated or original observations ranks? |
yes/no |
Corrections: Mann-Whitney, etc. as appropriate |
NON-NUMERICAL |
Observations cross-classified? |
Yes |
Two-Way Chi2 |
Finished |
|||||
No |
One-Way Chi2 |
Finished |
|||||||
NUMERICAL |
|||||||||
1 Group |
t for population mean |
Finished |
|||||||
2 Groups |
Observations paired? |
Yes |
Evaluated for Relationship? |
Yes |
t for pop corr coefficient |
Finished |
|||
|
No |
t for 2 pop means |
Condition 3 |
yes |
Corrections, then Finished |
||||
|
no |
Finished |
|||||||
|
No |
t for 2 population means |
Condition 3 |
yes |
Corrections, then Finished |
||||
|
no |
Finished |
|||||||
More Groups |
two factor classified? |
Yes |
Two-Way F |
Finished |
|||||
No |
One-Way F |
Condition 3 |
yes |
Corrections, then Finished |
|||||
no |
Finished |