Simply put, it was a mistake at both the tactical level, and with Japanese navy doctrine at the time. I would recommend you read "Shattered Sword" by Parshall and Tully. It's a fantastic book about the Battle of Midway, with lots of it from a Japanese perspective.

The Japanese search plan on the morning of June 4 involved seven float-planes, each assigned a heading to form a "fan" search pattern. The seven aircraft involved came both from carriers and from other ships. One B5N bomber each from the Kaga and Akagi, and then two floatplanes from the heavy cruiser Tone, three from the heavy cruiser Chikuma, and then one from the battleship Haruna.

This single-phase search pattern was fairly standard, and it was unusual to involve more aircraft for redundancy (it was only in mid-1943 when a two-phase search pattern became IJN doctrine). They could have used more aircraft from the carrier force, but that was against doctrine as it would have weakened the strike packages. Again, doctrine is the important point. The Japanese carrier force was seen as a very offensive weapon. In contrast, the Americans used over thirty reconnaisance aircraft during their sweeps in the battle. This is in part because of the fact that many of those aircraft were PBY flying boats based at midway, but also because American naval doctrine called for a relatively larger number of search aircraft. Finally, the Japanese floatplane (from Chikuma, I believe) whose track would have put it in a position to spot the American fleet flew above the clouds rather than below them, and missed it until it was on the return leg.

Again, I would really recommend that book, as it answers questions just like this and more. It explains this exact problem in far more detail than I could here on pages 107-113.