starch is used as the indicator in the determination of peroxide value in oil. when titrating against sodium thiosulphate the blue black colour is supposed to disappear.
Do You see the color change to brown when potassium iodide is added? This is caused by the formation of iodine (I2) which afterward is titrated with thiosulfate using starch as an indicator.
If not, the reaction conditions could be wrong. If I remember right, the the iodine-peroxide reaction should be performed in an acidic media.
It sounds a little odd. The change in color when iodide is added indicates that at least some of it is oxidized and addition of starch should then make it dark.
Maybe it is the test-conditions. Which standard method do you follow? AOCS Official Method Cd 8b-90 is widely used, but there are many others.