Test on Ordinal data with interval I have a query regarding statistical analysis. I have collected data on percentage of savings done by individuals. 
These data are ordinal, grouped into equal intervals with the exception of the final category representing all savings of 50% or higher. Because of this, I gather I can't apply a parametric test. 
I want to study whether there is significant difference in saving habits of two groups. Please suggest an appropriate test. I have a screen shot of the data.
Would a t-test or Mann Whitney U-test be appropriate?

 A: When the intervals are equally spaced, the data are perhaps ordinal by name only. The T-test summarizes the actual mean difference in % savings. Alas, the largest threshold group is not equal in spacing to the prior groups. So you must reframe the question slightly. 
If a non-parametric test is a consideration, then perhaps you are less interested in quantifying the actual % difference between men and women. In that case, you could use the T-test anyway and comment on the statistical significance as suggesting whether one group saves more or less than the other. This "non-parametric T test" is asymptotically equivalent to the Mann-Whitney test.
Another approach is to use ordinal logistic regression. This is also called a proportional odds model. Expand the data using a full data matrix of 252+98 observations or use a weighted dataset for the 10 tabular values:
Sex     Savings Weight
Male    0       100
Male    1       102
Male    2       ...
Male    3
Male    4
Female  0
Female  1
Female  2
Female  3
Female  4

Then fit the model with Savings as a response, sex as the main predictor, and supply the Weight to obtain a semi-parametric test of statistical significance.
Lastly, you are not forbidden from using a parametric model. Simply declare an underlying parametric model for the distribution of savings, and use expectation-maximization to estimate the latent beta (or other appropriate) distribution of savings in males and in females, then estimate the T-test of association.
