Student's t-table modification for the linear correlation coefficients estimation in the small samples cases

In the linear correlation coefficient calculations for the statistical significance estimation is often used the famous Chaddock's scale of the relationship between the studied phenomena with the characteristics like “weak”, “medium”, “visible”, “high”, “very high”, and for the significance evaluation used the Student's t-test table with the fixed alpha-level (
$\alpha=0.10;\ 0.05;\ 0.01$
) and with the available degrees of freedom. If the calculated values of the linear correlation coefficient are less than critical, then as usual the researchers will increase the number of initial observations N. However, in an unstable economics period this is not always possible. Therefore, we have the task of estimating the confidence interval for the calculated value of the linear correlation coefficient, especially to its lower bound (of confidence level): what if the calculated module of linear correlation coefficient will be met the reliability requirements according to the famous t-Student criterion? Moreover, this means that the significance in (both) cases is assessed on a step by step manner which is not fully expedient. For the decision of this problem we propose to use the modified Student's scale and then it is also possible to use the Chaddock's scale. As the raw data we use statistical aggregates with the limited size; after some modifications we create on this variation series and apply the consent criteria. For the noted problem solving we also must note the received equations for LCC error and non-strict inequality. Some part of the results was obtained with the program made in FoxPro 2.5 which was created by the member of the author's team (Anatoly Shikhalev).