Timeline for Statistical testing on non-random sample?
Current License: CC BY-SA 4.0
3 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Feb 19 at 20:53 | vote | accept | manaTEA | ||
| Feb 11 at 4:06 | comment | added | Glen_b | "But the validity of any generalization relies on your sample being random, or, at least, non-biased" -- unbiasedness (in a specific sense) may be ok for an estimate of some coefficient, but would not be sufficient for estimates of standard errors (and thereby for the usual forms of inference) to be correct. Statistical inference is based on a probabilistic model (e.g. on the basis of random sampling of a population or randomization to treatment). If the sampling or assignment to treatment is not random, then on what basis does the distribution of test statistics (or pivots in CIs) arise? | |
| Feb 10 at 16:48 | history | answered | Peter Flom | CC BY-SA 4.0 |