Barriers to Women Entrepreneurship. PLS vs. QCA: Do Different Methods Yield Different Results?

Building on research by Akehurst, Simarro and Mas-Tur (2012), this study analyzed internal and external factors in women entrepreneurship and linked these factors to the barriers that women face when starting businesses. To do so, two contrasting statistical techniques were used: PLS and QCA. Partial least squares (PLS) is an extension of principle component analysis (PCA) (Would et al., 1983). Both methods follow the principle known as soft modelling, which consists of analyzing models based on empirical data rather than on theoretical or logical constructions, as is the case in hard modelling (Sundbom, 1992). Therefore, PLS can be used to clarify complex patterns in the data (Semb, 2011).QCA is a set-theoretical method that assumes that the influence of certain elements on a specific outcome depends on combinations of these elements rather than the prevalence of the individual elements per se (Ragin, 2008). This method uses Boolean algebra to identify which combinations of properties are sufficient and/or necessary conditions to produce an outcome of interest (Fiss, 2007).This study had two objectives. The first objective was to extend the literature on barriers faced by women entrepreneurs, and the second objective was to observe differences between results of the same analysis conducted using two statistical methodologies: one quantitative (PLS) and one qualitative (QCA). After analyzing results from each of these techniques, we observed that family duties and difficulties in obtaining financing (both internal and external) were the main factors related to barriers faced by women entrepreneurs.