Why not simply use "calibrate" and "inaccuracy"? That gets the meaning across.

I have searched high and low for this and not come up with an established equivalent English term. And indeed, I don't think "pereza" is an established technical term in Spanish either; references are pretty few. I think it's a loose descriptive term.

I am very doubtful that it means "looseness" or "slackness". I can't see how the load reduction exercise that they propose in your text as a solution can eliminate looseness. If something is loose you solve that by tightening it. Loading the scale to full capacity and then reducing the load is surely going to make it looser, if anything, rather than tighter. I think the connection from "pereza" to "slack" and from "slack" to "loose" is probably a false one.

More likely, to me, is that "pereza" means that the scale is unresponsive or slow to respond.

One Spanish text that does give us a clue, I think, is this one (which I notice was quoted over on the Translator's Café discussion of this term):

"Cargar la balanza dos veces hasta el fondo de escala para eliminar la posible pereza o pegado de la "cruz"."
http://www.matematicasypoesia.com.es/metodos/mecweb08.htm

Over on TC someone said this was "looseness" and "sticking", but I don't think so; I think the two words refer to basically the same phenomenon, a "laziness" or "stickiness" in the mechanism that is actually loosened up by loading and unloading the scale a couple of times.

Actually this operation is mentioned in EN documents on calibrating scales, and it's called "exercising" the scale:

"In order to ensure that the scale is ready to operate over its full range of motion, the best practice is to exercise the scale prior to performing calibration. This involves loading up the scale with the maximum weight it can read and unloading it at least a couple of times."
http://www.google.com/patents/US7681431

As I say, I haven't been able to pin down a word commonly used for this phenomenon in English, and I've tried everything I can think of. There is "drift", which Phil mentioned, but that's an electrical or temperature-related malfunction, not a mechanical one (as this evidently is). And there's also "creep", but I don't think it's that either:

"CREEP
The change in load cell output occurring with time, while under load, and with all environmental conditions and other variables remaining constant; usually measured with Rated Load applied and expressed as a percent of Rated Output over a specific period of time."
http://www.ricelake.com/glossary.aspx

I think "sluggishness", which is close in meaning to "pereza", would express this. Exercising the scale and performing the load reduction exercise serve to eliminate sluggishness and stickiness in the operation of the scale. This other text also cited in the TC discussion suggests this to me too:

"Indicios de fricción en un manómetro son el desplazamiento "perezoso" de la aguja o movimientos erráticos de la misma."
http://www.equiposylaboratorio.com/sitio/contenidos_mo.php?i...

And here, albeit on a different kind of scale, is a use of "sluggish" which sounds to me as though it's describing a similar phenomenon:

"Friction at weighbeam bearings may reduce the sensitivity of the scale, cause sluggish weighbeam action and affect weighing accuracy."
https://bulk.resource.org/courts.gov/juris/j0642_27.sgml

--------------------------------------------------
Note added at 4 hrs (2013-09-13 20:24:47 GMT)
--------------------------------------------------

Sorry, what I've been referring to as "load reduction" is actually called the "decreasing load test".

An explanation