Many educational institutions such as mine at the University of New South Wales (UNSW) seem anti-collaborative in the assessment of students (also known as referenced marking). Marks are not pure marks, they are graded against peers in a student’s cohort to make a nice statistical bell curve or symmetrical normal distribution in marking as a type of scaling. Consequently, this feature can act as a disincentive for a student to help other students and even to hinder them. If most students do well in a course then the grades of most students can get scaled down. The better your mark, the more likely it is to be scaled down. The opposite is true if most students do poorly in a course. For educational institutions that use a referenced system, it can be seen that there is no incentive for collaboration. Discouraging collaboration is not the way. Collaboration should be encouraged because it is the essence of goodness and happiness. Educational institutions should be more collaborative. One way to do this is to reform educational assessment. Norm-referenced testing (the testing system that uses scaling and changing students’ raw marks relative to their cohort described above) is unlikely to aid collaborative education. Criterion referenced testing (which compares a student’s performance relative to explicitly defined goals) is better suited to do this. While educational institutions use a combination of assessment methods, my argument is that it is better to not have scaling if norm-referenced testing is used at all (indeed, it is better not to use norm-referenced testing altogether as it produces stress and anti-collaborative behaviour). “A house divided against itself cannot stand.” (Mark 3:24). If students, teachers, business people, citizens and family members are enemies of each other then educational institutions, workplaces, nations and families will collapse. If these same people are all friends then there will be world peace and harmony. For more information, see education assessment here, particularly norm-referenced testing here and criterion-referenced testing here.