Home – Home – Jornals – Contemporary Problems of Ecology 2006 number 2

The publication aims at attracting attention of Russian scientists to some methodological problems associated with both collection and analysis of the ecological data. By using recent publication by A. G. Kartashev & A. A. Kartasheva (Siberian Ecological Journal, 2004, 1, 25-34) I discuss the typical errors, such as (1) simple pseudoreplication, i. e. statistical conclusions based on evaluation units, not on independent experimental units, (2) multiple comparisons based on intercorrelated characters, including absence of correction for multiple comparisons in large tables of correlation coefficients, (3) absence of information on statistical distribution of the data and, consequently, potential problems associated with calculation of t-statistics and Pearson correlation coefficients for data showing non-normal distribution, (4) replacement of the values of correlation coefficients by their significance. The principal conclusion of the paper - impact of motorways on the structure of webs of the spider Larinioides patagiatus