Dear Statalists,

I have an outcome variable y, which records a company's transaction value. The variable shows "$1, $2 " till $10", "$10+, $20+" till "$100+", hereafter I drop the dollar sign, then 200+, 300+ till 1,000+; then 2,000+, 3,000+ till 10,000+; then 20000+, 30000+ till 90000+; then 100000+, 110000+, till 190000+; then 200,000+, 210,000+, till 290,000+. The maximum censored value is 9,500,000+. To sum up, below 10, it is not censored; above 10, it is censored depending on the value of y.

I want to estimate the effect of x variable on this censored outcome y. Usually, for censored data, we use the Tobit model. However, the standard Tobit model specifies one censoring criterion: e.g. either right-censored or left censored. I wondered what model I can use for this "multiple" censored data.

Thank you very much!
Best,
Kailin