This study, noting that most donors who have experience donating to universities are alumni and that alumni tend to identify with their alma mater, aims to analyze solicitation strategies that lead individuals to identify with nonprofit organizations and ultimately to examine the relative effects of the three dimensions of identification on loyalty. Specifically, the study first analyzes the relative influence of three dimensions of identification—cognitive identification, evaluative identification, and emotional identification—which are among the donation motives of university donors, on loyalty. Second, the study analyzes what solicitation strategies can elicit donor identification and what differences these solicitation strategies show across the three dimensions of identification. Based on these objectives, this study analyzed the historical donation data of donors to the university under study—specifically, individuals who had donated to a domestic Y University between March 1, 1996 and February 28, 2007. Using quota sampling based on this data, a sample of 252 individuals was drawn and a survey was conducted. Structural equation model analysis based on both objective historical donation data and survey results revealed that specific causes, emotional empathy, and interaction had a positive effect on emotional identification, while sentimental messages and personal relationships had a positive effect on cognitive identification. Furthermore, emotional identification had a greater influence on loyalty than cognitive identification and evaluative identification. The significance of this study lies in its multidimensional analysis of identification, a concept that has been relatively under-researched in the marketing domain, and in its analysis of the relative influence of each dimension on loyalty. Additionally, the study's significance can be found in its introduction of solicitation variables that have rarely been studied in the marketing domain, its advancement beyond existing studies that had been limited to experimental designs due to methodological constraints by conducting quantitative research, and its pursuit of model robustness by conducting surveys based on historical data of donors with actual donation experience and analyzing both sets of results together.