520 – Leveraging the American Community Survey as a Sampling Frame for the National Survey of College Graduates
A Review of Tests for Randomness in Time Series Data
Alicia Graziosi Strandberg
Temple University
Boris Iglewicz
Temple University
We review and compare a new proposed test with well known existing tests for detecting randomness in time series data, with special emphasis on stock market index data. By comparing popular variance ratio and traditional statistical tests plus a new proposed procedure, we have the most extensive simulation comparison of such tests. The investigated tests are compared over a diverse group of distributions, models, and stock market index applications. This study provides a useful guide to practical use of such testing for randomness procedures. We found that when evaluating common US stock market indices, the choice of data transformation can have a pronounced effect on test results.