2.6. Statistical methods

A series of binary logistic regression models were used to evaluate the association between TOXO seropositivity and dichotomous psychiatric diagnosis variables. A secondary series of binary logistic regression models assessing the same relationship between TOXO positivity and psychiatric diagnoses was also conducted adjusting for demographic variables.

The variables startle magnitude and latency were logarithmically transformed in order to account for a non-normal distribution. T-tests were conducted to assess the association between TOXO positivity and both acoustic startle magnitude and latency. A series of linear regression models were created to assess the association between TOXO positivity and either acoustic startle magnitude or latency, adjusting for demographics and psychiatric diagnoses. All regression models were assessed via backwards elimination in a stepwise manner. All statistics and data cleaning were completed using SAS 9.4 (Cary, North Carolina). Figures were created using Microsoft Office (2013, Redmond, Washington) and GraphPad Prism (La Jolla, CA).