As a trend is usually a linear fit through experimental values it does neither mean that there are no minor variations between the values nor that there is a constant rise between values but only that one can see a difference between earlier and later values of that measured property.
And besides all the answers on the actual influence of the longer alkyl chain on the acidity of the carboxylic group there are no big differences between the acidities for those small alkanoic acids between chain lengths of two to ten carbons.
However, there is a small overall trend $\left(+0.02~\mathrm{pK_a}\right)$ which can be seen in the following graph.
So one is a little bit stronger than the other but it is not very much.
One could start to argue why exactly those two differ "so much" in strength but looking at those nine values above and the "noise" around the trend I guess that makes no sense.
