We encounter dacitic/rhyolitic melts generally in explosive volcanic events. But some of them behave as effusive dome flow. Why do some silicic bodies behave effusively?
Explosive volcanic events happen mostly because of saturation in gases, and increase in pressure above the strength of the top of the magma chamber. As the magma decompresses or cools, H2O (mostly), CO2 (also important), and other gases such as H2S, SO2, HCl, etc, form bubbles and increase in volume. If the magma chamber is confined, there is also an increase in pressure.
An obvious corollary is that gases have to be present. If the magma is mostly dry and volatile-free, there are no volatile species to form bubbles and the magma cannot explode.
Another factor is the viscosity of the magma. The less viscous it is, the easier it is for the magma to flow (and not build up pressure) and also the easier it is for bubbles to escape the magma. Viscosity of magma depends on temperature (hotter = less viscous), composition (more alkalis = less viscous), crystal content (less crystals = less viscous).
Taken together, a hot alkali magma without crystals and no volatile species will flow, whereas a cold partly crystallised wet sub-alkali magma will probably explode, even if they have exactly the same SiO2 contents.
To complete Gimelist's answer: many silicic domes and flows lie right on top of a tephra layer with the same age and composition. See for instance this classic paper of Fink (1980) about Little Glass Mountain. This is interpreted as follow: a gas-rich magma ascends into a conduit and, because of its high volatile content, fragments, leading to an explosive eruption. The vent opening allows the rest of the magma to degas: it erupts effusively. Hence the lava flow or dome is emplaced on the tephra layer.
In 2011-2012, there was an eruption at Cordon Caulle (Chile) that showed that both dynamics can even happen at the same time! There was magma explosion and effusion from the same vent, simultaneously. This is explained by Castro et al. (2013) by a dyke geometry (figure 7) allowing some magma to ascend more slowly, thus having more time to degas. Schipper et al. (2013) also proposed a conduit model (figure 10) where shear zonation creates a network of permeable magma through which the explosive activity occurs.