Say we collected salary figures from our HR department and we knew to the exact dollar what everyone made this year and last year (interval/ratio data). We ran a parametric test (z or t) to see if there was a difference and our answer was significant at the 0.05 level. But let us assume that for some reason HR did not want to share exact dollar figures and was only willing to share categories with us (i.e., 50 people earned less than $20,000, 155 people earned $20,000 to 39,999, 250 people earned $40,000 to $59,999, 125 people earned $60,000 to $80,000, etc.) and we felt we needed to do a Chi Square test. How would be the significance?
To test if the salaries of different departments are the same using chi-square