The actual paper is short and easy to read. It’s based on photos posted online, so it says little on the pervasiveness of the behaviour. As the research puts it:
due to the often-contrasting colour of plastic relative to other anthropogenic or natural materials, plastic may be easier to spot for people, thereby influencing the incidence of reporting. Additionally, data collected through iEcology may be affected by potential biases arising from various factors that influence data collection among observers. These factors include individual and cultural subjectivity, such as the likelihood of posting unusual behaviour online.
The reasons for choosing artificial shells haven’t been tested either. The paper proposes some hypotheses (reduced weight, plastic emitting odours akin to actual seashells, lookin' sexy), but ultimately more research is needed to be able to affirm anything.