Language analysis of successful and failed startup founder interviews

This study examines the language employed in interview responses from both successful and failed startup founders. The analysis focused on identifying keywords, collocations, and comparing collocational patterns between the two groups. The corpus comprised 40 interviews: 20 from successful startup founders and 20 from failed startup founders. These responses were sourced from an online platform and purposefully selected based on three main criteria: (1) the country of origin of the startup, (2) the startup category, and (3) the length of the interview. Data cleansing was conducted to ensure comparability. In particular, negative questions and their corresponding answers were excluded, as such content was not uniformly represented across both groups. Retaining these inconsistently occurring elements would have introduced bias and limited the reliability of direct linguistic comparisons. Keywords and collocations were identified using AntConc software and subsequently categorized according to a predefined analytical framework. The findings reveal that successful startup founders tend to use more specific and targeted collocations, whereas failed founders typically employ broader and more generalized collocational patterns.