Results of the ordinal regressions on aggression towards familiar people. The graph represents the estimates (β) and standard errors (s.e.) for the significant models explaining the variance in the aggression towards familiar people during COVID-19 confinement. The dotted line represents the intercept, the geometrical figures represent the estimates and the coloured shadows represent the standard errors. When the response options were categorical, rather than binary or continuous, the reported parameters are compared against the reference parameter for each variable: intact female for the sex of the cat, some are in lockdown and some can go to work for the lockdown of the household, adults without children for the household role, employee for the job, degree for the education, female for the gender of the owner. Significant comparisons are indicated with an asterisk. Positive estimate values are on the right, whereas negative estimate values are on the left.