Hi Chaudry! Yes, sometimes a p value less than 0.10 can be considered significant. However, p value is not enough to validate a hypothesis in PLS-SEM. You must assess the prediction relevance of the independent variable, depending on the objective of your research. I recommend you to check the value of the path coefficient, the confidence intervals (bias corrected), R square, and f square value too. After assessing all these measures you can validate if the independent construct is relevant for the prediction of the dependent variable or not.

The choice of an alpha level to determine what p-value to consider "significant" is entirely arbitrary. The .05 is (merely) conventional in many disciplines; a stricter .01 is common in others. If your sample is only modest in size, or the research is in a new area without much prior literature, you could argue for a more lenient standard of .1, to reduce the risk of Type 2 error (failing to reject a false null hypothesis).

Note too, most software generates 2-tailed p values. If that's the case here, and if the sign(s) of the effect(s) in the hypothesis were correctly predicted in advance by your theoretical model, then you could justifiably use a 1-tailed test, with a p-value of .057/2 = .0285. Then you could report the result as "significant at the .05 level in a 1-tailed test."

The threshold for significance depends on the sensitivity and importance of the subject being analyzed. It is advisable to consult relevant literature for guidance specific to your field.

In reporting, you may note that a result is significant at the 0.10 level but not at the stricter 0.05 level, if applicable. To clearly communicate significance levels, you can adopt the following notation:

• *** for significance at the 0.01 level
• ** for significance at the 0.05 level
• * for significance at the 0.10 level and
• No symbol for results that are not significant at the 0.10 level