Report on the Second Random Sampling Survey for Estimating the Outturn of Wheat in the United Provinces 1944-45

Again whereas 'last year the length of the plot was marked along the direction of the rows in the field this year the instructions provided that in view of the difficulty of identfying the direction of the rows experienced last year it should be marked along the direction of the length of the field. [...] Considering that the assumption regarding the proportional distribution of the villages accoring to the area under wheat in the different tahsils and the homogenity of the average yield and the sampling error in all tahsils are not fully realised in practice the agreement between the actual number of villageand the estimated one must be considered to be satisfactory. [...] The first source of bias may necessitate correction in the estimated average yield in the event of an association between the area of a village (or field) and the yield per acre and the second in the event of an assocition between the position of a plot in the field and the plot yield. [...] The relative position for the length of a plot is defined by the ratio of the random inkinber for length for the starting point of a plot to the length of the field minus the length of the plot and is given in m4111111'1. [...] It is partly due to the fact that in usim the two-digit list of random numbers for the location of plots which was the list most often consulted all the qanungos started with the first number of the first column with the result that the characterisCes of the first column got repeated in the selection.