Why is chance such a threat to a political scientist's hypothesis?

I need you guys to tell me if this is right?

Chance is on the political scientist's side until up to a certain point. He makes a hypothesis and sends out his colleagues to do the research. If the research comes back with less than five percent, then the result of the study was by mere chance and the political scientist would be face with "what ifs". If it is more than five percent, then the result is statistically significant, and therefore, the result was not by chance. The best way to eliminate that what if factor, which chance contributes to the research, is by using random sampling in order to gain a statistic of what is being researched. That way, the
political scientist would, for example, get a good idea of a significant portion of the population's interest or knowledge about politics for his research. Suppose he hypothesizes that older Americans with better education would be most interested in politics and would therefore be more likely to vote, or get involved in political demonstrations, because another hypothesis would be that those with better education
are somewhat or completely dissatisfied with the way the government is handling the country's issues. Random sampling would give the political scientist a statistic that shows whether the result is by chance or not.

Don't forget that, even if the test statistic is significant, it does not rule out the fact that the results were gotten by chance. There are always what ifs, that is what the threat is.