Is there solid research comparing learning loss I. states where schools were closed longest had greater loss than those closed shortest?