There are millions of MOOC participants who vary in gender, age, educational level, employment status, intentions, etc. Although MOOC participants' characteristics have been studied, there is still a lack of knowledge of the divergence between the participants and completers of MOOCs with different levels of difficulty. The term 'level of difficulty' as used in this paper encompasses, besides the difficulty of covered topics, the variety of supportive teaching methods and different course durations. The aim of this study was to determine the demographic and social background characteristics of participants and completers in three programming MOOCs with different difficulty levels. It was found that the difficulty of a topic is related to gender, age and educational level distribution in MOOCs. According to our results, previous experience in the topic and the difficulty level of the MOOC influence completion. However, our results were less clear-cut regarding the correlation of age, education and employment status with difficulty level of MOOC. The results can be useful for MOOC instructors in supporting different participant groups, for example, by allowing more flexibility for specific participant groups.