SOME ASPECTS OF THE DESIGN, ANALYSIS AND PRESENTATION 
OF PHOTOGRAMMETRIC EXPERIMENTS *) 
by A. M. Wasser, 
The Survey of Egypt, Giza, Egypt. 
‘It is much easier to be 
critical than to be correct? 
Disraeli 
Abstract. 
The limitations of the reliability of inference from some published photo- 
grammetric experiments are discussed to illustrate shortcomings of design and 
inadequacy of published accounts. 
The discussion leads to advocate a reversal of the current approach to photo- 
grammetric experimentation by showing the advantages of starting with a frame- 
work of primary experiments which reveal the suitable conditions for more 
detailed investigations of the relevant factors. The need for this approach emerges 
from the largeness of the number of factors which influence the accuracy and 
economy of air survey, and also from the possible entanglement of their effects. 
This approach owes its practicability to the recent advances in the science of 
statistics; and has proved invaluable in many fields of industrial research. 
The analysis of the reviewed experiments, which are drawn from work in 
photographic distortion, instrumental errors and air triangulation, demonstrates 
some of the statistical techniques, and explains the nature of inference. 
Approaching experimental photogrammetry along these lines meets with some 
difficulties due to the subjectivity of the attribute of ‘goodness’, for instance in 
connection with photographic quality. It will be seen that such difficulties can be 
tackled along the same lines; and it is hoped that a wider appreciation of the 
power of the statistical approach to the planning of experimental investigations 
will help secure the necessary facilities. 
f. 
This paper was written to draw attention to the need for a wider application 
of the recent statistical principles of experimental design and analysis. 
Statistics is of course no novelty to the surveyor: The theory of errors and 
the method of least squares, which are at its foundation, are no strangers to him, 
and have served him well. 
But the survey problems which need modern statistics most, have arisen from 
the application of photography. Ground photogrammetry did not create serious 
problems. The intricate problems of photogrammetric indeterminism started piling 
up the moment the camera deserted the theodolite and took wings. 
The determination of cameras to fly higher and faster, and their obstinate 
defiance of stabilization, are as manifest as the surveyors’ eagerness to be relieved 
of all ground work. But the desires are incompatible. 
1) This paper will be read before the Seventh International Congress of Photogrammetry, Wash- 
ington, D.C., 1952. : 
8