Assuming that you have basic concept of simple linear regression model

so as in simple linear regression model we have dependent variable (or response variable) is always a continuous variable like weight height, or any thing which is measured in ratio scale, and independent variable (factors effecting response) which may be continuous, or binary variable (like gender, or any variable with response yes no) so we use simple or multiple regression model (extension of simple linear model with more independent variable).

now if you have a situation in which your response variable is binary then we use probit model.

as binary variable response is only of two categories so to use in the analysis we represents them with 0, 1. 0 mean absence and 1 mean presence . e.g. if we define a variable as Male then 1 mean Male 0 mean female and so on.

now on the basis of the model when we have to tell about the response variable which should be between 0 and 1, so we incorporate a probability distribution (Normal) to get predicted values between 0 and 1. as a result this predicted value is interpreted as probability.

let say we have a certain levels of treatment which we apply to a some fishes, to check on what level they begin to die. we define a variable die with 1 mean fish dies, and 0 mean survived, with different level of treatment as independent variable.

so when we fit the model then it will tell the probability of dieing of fishes for a certain level of treatment

Assuming that you have basic concept of simple linear regression model

so as in simple linear regression model we have dependent variable (or response variable) is always a continuous variable like weight height, or any thing which is measured in ratio scale, and independent variable (factors effecting response) which may be continuous, or binary variable (like gender, or any variable with response yes no) so we use simple or multiple regression model (extension of simple linear model with more independent variable).

now if you have a situation in which your response variable is binary then we use probit model.

as binary variable response is only of two categories so to use in the analysis we represents them with 0, 1. 0 mean absence and 1 mean presence . e.g. if we define a variable as Male then 1 mean Male 0 mean female and so on.

now on the basis of the model when we have to tell about the response variable which should be between 0 and 1, so we incorporate a probability distribution (Normal) to get predicted values between 0 and 1. as a result this predicted value is interpreted as probability.

let say we have a certain levels of treatment which we apply to a some fishes, to check on what level they begin to die. we define a variable die with 1 mean fish dies, and 0 mean survived, with different level of treatment as independent variable.

so when we fit the model then it will tell the probability of dieing of fishes for a certain level of treatment

This is in continuation to the above answer from statistical perspective.

Probit is a statistical tool to analyze the quantal responses (all or none) to stimuli (dose). It could have applications in several areas like toxicology etc.

Probit analysis is routinely used in potency assays for many antigens/vaccines like Diphtheria toxoid, tetanus toxoid etc.The measure of activity of the antigen is death/survival when the immunized animal is challenged with respective toxin. Typically in these assays set groups of animals are injected with different doses of a reference standard with a known unitage and the test sample (unknown) .Animals are given time to mount immune responses and subsequently challenged with toxin.If the responses in the reference standard and the test sample are linear and parallel by the probit analysis, the potency of the sample can be determined.

Also find enclosed, an article from the web, which to me seems to be fairly uncomplicated for people with minimum statistical background, to understand.

This is in continuation to the above answer from statistical perspective.

Probit is a statistical tool to analyze the quantal responses (all or none) to stimuli (dose). It could have applications in several areas like toxicology etc.

Probit analysis is routinely used in potency assays for many antigens/vaccines like Diphtheria toxoid, tetanus toxoid etc.The measure of activity of the antigen is death/survival when the immunized animal is challenged with respective toxin. Typically in these assays set groups of animals are injected with different doses of a reference standard with a known unitage and the test sample (unknown) .Animals are given time to mount immune responses and subsequently challenged with toxin.If the responses in the reference standard and the test sample are linear and parallel by the probit analysis, the potency of the sample can be determined.

Also find enclosed, an article from the web, which to me seems to be fairly uncomplicated for people with minimum statistical background, to understand.

Thanks a lot ppl..

Thanks a lot for your responses. They are really helpful. Maybe, more examples in social sciences would also do.