The development of adequate collocational knowledge is important for foreign language learners; nonetheless, learners often have difficulties in producing proper collocations in the target language. Among the various ways of learning collocations, the DDL (data-driven learning) approach encourages independent learning of collocations and allows learners to directly use corpora and tools to search for proper collocations. Concordancers have been one of the key tools in the DDL approach. However, based on the findings of empirical studies, many existing web concordancers have not been very helpful in retrieving collocations. To help EFL students and teachers quickly locate proper collocates, this article introduces a web-based collocation retrieval tool, WebCollocate, which is based on a large part of speech-tagged Gutenberg corpus. To determine if the new tool can facilitate the searching of collocations, this tool and the Hong Kong Polytechnic web concordancer were used by two groups of college EFL students to find proper collocates in a translation task. The results showed that the students who used the WebCollocate tool found more proper English collocates. In addition, a group of 35 pre-service English teachers were also invited to evaluate the effectiveness of this collocation tool. These EFL teachers indicated that they could easily find proper English collocates with the help of WebCollocate. However, they also suggested that the search options and the presentation of search outcomes of this tool can be further improved. It seems that a more dedicated collocation retrieval tool can facilitate collocation teaching and learning, but the functions of this tool can be further enhanced. The findings of this study can be useful for language teachers, researchers, and developers of corpus-based tools.[Copyright of Routledge. Full article may be available at the publisher's website: http://dx.doi.org/10.1080/09588221.2010.526945]