Problem of the First Person

Many writers and teachers believe that it is never appropriate to use I or we in professional scientific and technical prose. Some science writers go so far as to avoid even third-person agent-action structures, writing "In the classic research, X was found by Smith and Jones" rather than "In their classic research, Smith and Jones found X." In both cases those writers are mistaken. Good science writers use the first person all the time, and there is no reason in the world to avoid third-person agent-actions structures.

Good science writers who use the first person follow this general pattern.

1. When the action is one that only the author can perform, actions such as state, study, conclude, decide, etc., then use the first person or a disguised first person.

Not: Substantial agreement with the classical analysis was found in [the authors'] previous studies.

But: In previous studies, we found substantial agreement with the classical analysis.

Or: In their previous studies, the authors found substantial agreement with the classical analysis.

Not: The conclusion that LDGB8 is not one of the affected structures must therefore be reached.

But: We must therefore conclude that LDGB8 is not one of the affected structures.
2. When the action is one that anyone who repeated the research could perform, actions such a measuring, calculating, testing, evaluating, etc., you can (i) use the first person if you do so rarely and do not focus too much on yourself, (ii) use a general agent (researchers, engineers, etc.), or (iii) use an agentless passive.

Notice that the kinds of actions that call for the first person tend to be concentrated at the beginnings and ends of articles and technical reports. Those are the places where an author either uses metadiscourse to set up or comment on the text or focuses the reader's attention on what is original and important about the author's research. The actions that call for agentless sentences are concentrated in middles, where the presumed objectivity of the scientific method should predominate. 

You should no more avoid first person constructions than you should avoid passive constructions. Both have their uses. Your job is to understand their uses and to use them when they are appropriate. Note how these two passages use both active and passive constructions to tell their stories.

Moe has stressed that the surgeon must accurately measure the curve of the spine and analyze levels of rotation. Moe also stresses that in order to determine the flexibility of the lumbar curve, the surgeon must use preoperative supine side-bending roentgenograms. He advocates that the thoracic curve be fused from the superior neutrally rotated vertebra to the inferior neutrally rotated vertebra. If a thoracic and lumbar curve are combined and the lumbar curve on side-bending has been corrected to equal or exceed the thoracic curve, then Moe advocates fusing only the thoracic curve.

Following the rules of strain strengthening, we can predict mechanical properties (both tensile and compressive) in any direction or location in a part formed by any of the basic deformation processes. These rules incorporate the strength designation already explained, the uniaxial plastic stress/strain characteristics of the material, and the strain history induced by the forming process.

These rules were developed in extensive research into methods for characterizing materials property and plastic deformation. The deformation processes that were experimentally and analytically studied included the cyclic axial deformation of cylinders, cyclic deformation of cubes in three perpendicular directions, bending and unbending of flat specimens, cyclic torsional deformation of cylinders, shearing of blanks, deep drawing of channel sections and cylindrical cups, bar drawing, forward and back extrusion, and cold rolling.

One basis for calculating the strength of a formed part is the uniaxial stress/strain relationship of the original material, which can be determined by a tensile or compressive test. The calculations determine the plastic behavior of the material by using the exponential relationship,

s = so em

where s is the stress associated with a strain, e, and so and m are the stress coefficient and strain exponent, respectively.

A second requirement is that we know the strain history at the critical location in the formed part. For the six basic deformation modes, the strain history can be determined analytically. However, for complex shapes made by two or more basic deformation processes, the engineer must obtain experimental data from the shop floor. Most formed parts undergo more than one cycle of strain when they are fabricated. For example, the metal may first be stretched and then later compressed. Sometimes, three or four such cycles can occur.