Although there is a large body of literature on the use of connectors by EFL learners in argumentative writing, African language learner groups are not represented and most studies have adopted a cross-sectional approach. This paper seeks to analyze how frequently connectors are used on the basis of pseudo-longitudinal data from first, second, and third-year undergraduate students from the Institute for Applied Pedagogy, University of Burundi. The study adopted the corpus linguistic and quantitative approaches. The data were collected through 2 writing assignments given to students in 2019. In total, 622 essays, amounting to 306,664-word tokens, were collected. Data preparation followed three steps, namely, typing the handwritten essays, preparing the TXT files, and metadata for the corpus. The latter was searched using AntConc tool, on the basis of a comprehensive list of 95 connectors, classified into 6 categories. The tool availed frequency information, that is, how often each connector occurs in the corpus. Percentages were computed using XLS facility and are presented in tables and graphs. Results indicate that 19 connectors were frequently used by learners, the most frequent ones being those which add information followed by those expressing a result relationship. However, connectors expressing transition and summation were found to be rarely used in the corpus. Another discovery is that learners in all three years of study seem to have the same preferences in terms of connectors. Furthermore, cases of misuse such as wrong combinations and incorrect spelling, semantic, stylistic and syntactic errors were observed at all levels. The findings of this study can be useful to learners, teachers, and material designers.