In recent years, it has become more and more commonplace for investment performance attribution analyses to be carried out with a daily observation periodicity.  The justification for changing to daily observation frequency from longer periods (such as months) is that these analyses are believed to be more able to accurately reflect the actual investment returns on a fund.  Such beliefs are based on a series of operational, mathematical and statistical assumptions that are demonstrably false.  In actuality, applying typical  attribution methods to daily data leads to analytical conclusions that are highly biased and unreliable.  This study will examine the logic behind these assumptions and review the existent finance literature.