Viewing page 20 of 152

This transcription has been completed. Contact us with corrections.

March 20, 1930.

White House Conference Committee on Growth and Development.

Sub-committee on Graphs and Tables

E.B. Wilson, C.R. Bardeen, R.E. Scammon

As most of the graphical and tabular material for the reports will be collected from the literature, it will be in very variable form. Time and money will probably both be lacking to regraph and retabulate the material. All authors will, however, do a great deal to render graphs and tables useful instead of useless if they will give thoughtful attention to details necessary to rendering the material precisely defined and reasonably comparable. Some suggestions follow:

(1) State specifically what are the units, metric or English, in each table or on each graph.

(2) Avoid ruling and boxing in tables.

(3) Specify carefully what the entries in the different columns mean. Headings over the columns may be quite insufficient to do this; the columns may be numbered and a legend furnished for the table as a key to the numbers.

(4) Avoid useless sub-tabulation; stick to the main thing.

(5) State total number of observations.

(6) Divide the sexes or specify that the table is mixed. Separate white from colored or specify that the table is mixed. (In cases where the table is mixed it might be useful to have some estimate of the percentages of the two sexes or of white and colored - the original material may, however, not afford such information.)

(7) Divide by ages, preferably by months in the first year of life and by years thereafter. In case of age groups, indicate clearly what are the limits of age. Thus "Age 5" means sometimes 4 1/2 to 5 1/2, though generally 5-6. An age group 10-14 may mean 10 up to 14 though generally 10 through 14.

(8) When giving mean values state if possible the probable error or the standard deviation of the mean and specify which it is, as 7.231+/-.16 (p.e.) or 7.23+/-.24 (s.d.).

(9)In drawing and lettering figures bear in mind how they will look when reduced to page width or to the size specified.

(10) Design figures so as to get in the essential elements legibly without waste of space. Mark scales carefully and furnish a legend that will make the figure intelligible by itself.

(11) Consider the possibility of grouping simple figures together so as to be reduced together in one block. But avoid overcrowding any figure with a series of lines (full, dotted, dashed, etc.,) until it is difficult to decipher.

(12) In the main, plot histograms rather than frequency curves; they are more likely to be comparable, particularly when the basal age-groupings are different. When age is the base, space time out uniformly instead of irregularly. (It may be advisable to have a special figure for the first year by months.)