Barely Significant

Damaged p-values

Two kinds of p-value that cannot be taken at face value. Neither is quietly discarded, and neither is allowed into a median or a histogram anywhere else on this site.

The missing decimal point

3 sentences report a p-value whose decimal point did not survive: p <01 for p < .01, p = 06 for p = .06. The tell is a leading zero on a value with no separator at all, which describes a p-value and almost nothing else.

These are read as the author meant them, because the alternative is worse: taken literally, p <01 is p = 1.0, which would sit in the ordinary bands looking like a real result.

Not a probability at all

A p-value is a probability. It cannot be exactly zero, it cannot be negative, and it cannot exceed one. These 10 sentences report one that is.

Most are p = 0.000, which is a rounding decision printed as a certainty: the honest form is p < 0.001.

slightly significantP = 4.9probabilities cannot exceed 1
We went on to show that the use of exogenous normalizers only for absolute RT‐qPCR concentrations measurements did not influence the correlation with microarray data (standard normalization) for miR‐16‐5p (r = 0.52; P = 7.1E‐2) and miR‐486‐5p (r = 0.84; P = 2.9E‐4), although the correlation becomes slightly significant for miR‐93‐5p (r = 0.55; P = 4.9E‐2) ( Figure 4 middle panel).
only marginally significantP = 0.000a p-value is never exactly 0
Some studies showed significant differences ( P = 0.09,[ 66 ] P = 0.063,[ 146 ] P = 0.054,[ 31 ] which were only marginally significant, P = 0.000,[ 71 ] P = 0.000,[ 144 ] P = 0.003,[ 65 ] P = 0.000,[ 4 ] P = 0.000[ 70 ]) and some did not demonstrate significant or marginally significant differences ( P > 0.1).[ 66 , 148 , 149 , 150 , 176 ] TREATMENT OF CONGENITALLY MISSING TEETH CMT has direct clinical implications.