As you have two groups and you want to compare the means of the two groups, you can use an independent t-test. An independent t-test is a parametric test that is used to compare the means of two independent groups.

It's too small a group to draw a conclusion, even if you apply a test.

Muneeb Shah is correct when he said that the mean difference between the groups is tested using a t-test. However, this test is based on some assumptions about the distribution of the variable, particularly that the distribution is normal and that variance is constant (equal in all groups). Using these assumptions makes the test powerful, but it make the test also vulnerable when these assumptions are not sensible for the variable you are studying. So you must check if these assumption make sense in your case, and what you can do if they don't make sense.

"Non-parametric tests" (aka rank-based) make different assumptions and test different hypotheses - not necessarily the mean difference in which you are interested. Additionally, rank.based tests have very little power in small samples, so these are not a good choice in your case.

"Bootrstapping" is another way to find the statistical significance of a mean difference (or any other statistic you may be interested in). In contrast to rank-based tests they allow you to test the desired hypothesis using fewer assumptions than the t-test, but like rank-based tests they have virtually no power in small samples.

Note that any test of statistical significance is a "data sufficiency test". It tells you if the information provided by the data about a parameter of interest (here: the mean difference) is sufficient to statistically discern it from a hypothetical value. Therefore I disagree with Jaspriya Sandhu's response.

Bootstrapping; perhaps a robust one 💞

Jaspriya Sandhu is right. With possibly very unusual exceptions, when the variability is essentially zero, 4 per group is just too few to test for significance. Of course you can and should summarize the data.

One sample t-test