Published 1996

3 Farmer-Based Assessment of Soil

Quality: A Soil Health Scorecard1

Douglas E. Romig, M. Jason Garlynd, and Robin F. Harris

University of Wisconsin
Madison, Wisconsin

Soil quality is undergoing a dramatic process of redefinition. Its meaning has

moved beyond soil productivity to encompass environmental quality, food safe-
ty, and animal and human health (Parr et al., 1992; Doran & Parkin, 1994; Karlen
et al., 1997). Farmers also see the shortcomings of the old definition that limits
soil quality to yield potential and nutrient levels. In listening meetings held in
Wisconsin, many growers felt that the biological health of the soil was not receiv-
ing scientific attention (Kelling, 1989).
Soil health, a more integrative term preferred by some farmers to soil qual-
ity, is recognized and assessed by farmers using indicator properties of both soil
and nonsoil target systems (Harris & Bezdicek, 1994). While a number of ana-
lytical parameters to measure soil quality have been proposed by scientists (Lar-
son & Pierce, 1991; Karlen & Stott, 1994; Doran & Parkin, 1994), farmers' diag-
nosis of a soil's condition primarily uses qualitative or sensory means in addition
to quantitative data (Romig et al., 1995). Granatstein and Bezdicek (1992) sug-
gest that the tools or indices developed to measure soil quality, be they for
research, farm, or regulatory purposes, need to integrate both sensory and quan-
titative aspects.
Descriptive and integrative approaches used by farmers to characterize soil
health provide a mechanism for field assessment and monitoring of soil quality
by scientists and farmers. Possible descriptive indicators to characterize and mon-
itor soil quality are given by Arshad and Coen (1992) and Reganold et al. (1993),
including surface crusting, evidence of erosion, ponding of water, vegetative
cover, soil structure, friability, and consistence. Furthermore, the effectiveness of
qualitative soil survey techniques has been demonstrated for describing changes
in soil surface properties (Grossman & Pringle, 1987), and evaluating land suit-
ability with descriptive indicators (Sharman, 1989).

1
The Wisconsin Soil Health Program has been supported by the University of Wisconsin's Cen¬
ter for Integrated Agricultural Systems, and the Agricultural Technology and Family Farm Institute;
the Wisconsin Department of Agriculture, Trade, and Consumer Protection's Sustainable Agriculture
Program; the Wisconsin Fertilizer Research Council; the Wisconsin Liming Materials Research
Council; and the Kellogg Foundation through the Wisconsin Integrated Cropping System Trial.

Copyright © 1996 Soil Science Society of America, 677 S. Segoe Rd., Madison, WI 53711, USA.
Methods for Assessing Soil Quality, SSSA Special Publication 49.

39
40 ROMIG ET AL.

Several authors have promoted the development of various analytical tools

and indices to assess soil health and quality that would document changes that
occur due to management (Haberern, 1992; Granatstein & Bezdicek, 1992;
Karlen et al., 1994; Doran & Parkin, 1994). Assessment tools and strategies need
to serve not only scientists and resource managers, but also should be measured
by how well they meet the needs of farmers and enhance their analyses (Thrupp,
1989; Chambers, 1993; Acton & Padbury, 1993). To meet the challenge and
develop a practical index for soil quality and health, the University of Wiscon-
sin's Soil Health program has promoted a partnership between farmers and sci-
entists. This chapter presents the outcome of structured farmer interviews, the
Wisconsin Soil Health Scorecard, (hereafter referred to as the scorecard), an
assessment tool for soil quality based on farmer knowledge of soil health. The
scorecard is presented in full immediately following the explanatory text. Details
concerning the methodological approaches are presented in Romig (1995) and
interpretation of farmer knowledge of soil health is amplified in Romig et al.
(1995).

INTERVIEW METHOD

A qualitative approach was taken to solicit and analyze farmers' knowledge

of soil health through structured interviews. These interviews were given to 28
farmers in the summer of 1993, and explored the question: "How do you recog-
nize a healthy soil?" Participants were associated with the Wisconsin Integrated
Cropping Systems Trial (Cunningham et al., 1992) and operated conventional
and low-input cash grain and dairy farms typical of southeastern Wisconsin.
Agricultural soils of the region are formed in silt overlying glacial till or outwash.
Use of an interview guide ensured that each soil health interview covered
the same material (Garlynd et al., 1994). Interviews were conducted on the farm
and recorded on tape for further analysis. The soil health interviews were con-
ducted in two stages. First, open-ended questions were used to explore how farm-
ers recognize a healthy soil. This allowed for free association and brought to light
characteristics of a healthy soil that were in the forefront of the farmer's mind.
When it seemed that the participant had addressed all the indicator properties,
questions became closed-ended. In the second phase of the interview, farmers
were asked to consider if a soil's health was related to specific properties that had
been collated in earlier research (Garlynd et al., 1994), but the farmer had not
addressed in the first phase. Throughout the entire interview, probing questions
were asked to gather descriptions of each soil health indicator, be they words,
phrases or numerical values.
The interview responses were coded and these data were analyzed to deter-
mine the most important properties of soil health as perceived by farmers. Prop-
erties were considered greater in importance if they were: (i) used by the major-
ity of the farmers, (ii) mentioned earlier and more frequently, and (iii) mentioned
in the open-ended question period rather than being prompted by closed-ended
questions. A procedure was developed based on these assumptions to rank the
properties relative to one another and determine which indicator would be includ-
FARMER-BASED ASSESSMENT 41

ed in the scorecard (Romig, 1995). Additionally, all descriptions of healthy and

unhealthy soil for each indicator property were cataloged.

A SOIL HEALTH SCORECARD

Conceptual Basis
The interview analysis created a database of farmer knowledge about soil
health and quality, identifying, and characterizing important soil health proper-
ties. Transformation of these data into scaled items and their integration into an
assessment tool for soil health introduces a measure of subjectivity. Indeed, any
integration of data, be they analytical or descriptive, will implicitly confer a qual-
itative nature to an index (Granatstein & Bezdicek, 1992). To minimize the level
of subjectivity, literature concerning human health and behavioral sciences was
reviewed to define health and deal with issues of scaling.
Overviews of health definitions are provided by Larson (1991) and Bowl-
ing (1991). By far the most common and simplest definition of health is the med-
ical model—the absence of disease and disability. There are shortcomings of the
medical model, namely one can feel ill without having a disease and, similarly,
one can be in good health even though a disease has been diagnosed. More holis-
tic models extend to mental, social, and environmental aspects of health and often
include positive dimensions: functional fitness, wellness, and quality-of-life (Lar-
son, 1991; Bowling, 1991). Culyer (1983) cuts across many of the abstract dif-
ferences in various health models to a pragmatic definition in terms of a set of rel-
evant characteristics. Health indicators in this characteristics approach include
measures of functional capacity (ability to perform normal tasks and activities),
morbidity (illness), and vital statistics (attributes of a particular health state).
Through conversations with farmers, a set of soil health indicator proper-
ties was identified for soil health characterization. These included both the capac-
ity of a soil to perform certain functions (infiltrate, decompose, cycle nutrients,
and others) and attributes or vital statistics (soil color and structure, root mor-
phology, earthworms, animal health, and others). Employing the characteristics
definition of health (Culyer, 1983) allows soil health to be described and evalu-
ated with respect to both functional ability and structures or attributes. In the
scorecard, each soil health indicator is operationalized to conform to the follow-
ing subjective rating scale:
1. Healthy: Performance of function is optimal and structure is normal.
2. Impaired: An abnormality in function and/or structure.
3. Unhealthy: Severe restriction or inability to perform function considered
normal, severe deformity or loss of structure, disabled.
The scorecard's soil health scale rests on the assumption that the majority
of indicator properties are subjectively measured by the senses (look, smell, or
feel). Indicators of soil health vary in quality or character rather than in an spe-
cific measurable quantity. The scale used in the scorecard does not measure the
exact magnitude of difference among healthy, impaired, and unhealthy cate-
gories, meaning the scale is at the ordinal-level of measurement.
42 ROMIG ET AL.

Scaling methods for objects that have degrees of magnitude that are sub-
jective or qualitative are addressed by psychologists Stevens (1975) and Torger-
son (1958) who studied the characteristics of humans as measuring instruments.
Their research demonstrated that humans can make accurate and internally con-
sistent judgments of phenomena and partition stimuli into ordered, homogenous
classes. In short, an individual has a relatively constant sensitivity to differences
even when there is no objective, physical scale by which they would measure.
Rating scales initially produced by people are at an ordinal level of measurement.
That is, objects are ranked along a scale as to the degree they possess a given
characteristic without a distinct and uniform distance between each class or cat-
egory (Blalock, 1972). An ordinal scale can be mathematically transformed into
interval scale values through various psychological scaling procedures, given
enough data from trials that test the rating scale. Brown and Daniel (1990) review
several of these scale transformation methods and demonstrate their application
to the development of environmental rating scales. Psychological scaling tech-
niques also have been widely used to develop rating scales for human health
indices (Rosser, 1983; McDowell & Newell, 1987). Given the quantity of data
necessary to construct interval scales from ordinal categories, and the limited
empirical data that have been collected with the scorecard to date, the scale at
which soil health indicator properties are measure by the scorecard remains at the
ordinal level.
There are, however, simple techniques, such as partition or category scal-
ing suitable to assign numerical values to equal appearing intervals linearly relat-
ed to stimuli measured on a subjective scale of magnitude (Stevens, 1960; Torg-
erson, 1958). Farmers described soil health properties on an essentially dichoto-
mous scale, providing qualitative descriptions of healthy or unhealthy extremes.
Table 3–1 is a catalog of representative terms farmers used to describe the top 20
soil health indicators at healthy and unhealthy states, making these end-points rel-
atively well defined. By employing a partitioning scaling method, descriptions of
the impaired state of health for each indicator property were interpolated to give
equal-appearing intervals between healthy and unhealthy. The scorecard pre-
views the descriptive range of the indicator property and anchors the rating crite-
ria to a numerical scale for unhealthy (0 to 1), impaired (1.5 to 2.5), and healthy
(3 to 4) options. The numerical rating scale allows the scorecard user to grade
each indicator with some degree of sensitivity. Yet it is important to remain cog-
nizant that this rating scale is ordinal, thereby limiting the mathematical opera-
tions the data generated by the scorecard can support (Blalock, 1972).
Every attempt was made to keep indicator property descriptions used in the
scorecard aligned with the language and intent of the farmers who were inter-
viewed; however, some indicators (hardness and color) employ soil science tech-
niques used to describe morphological properties. In other instances, statements
regarding analytical features (yield, organic matter, and pH) were reviewed by
extension soil scientists. Users also will note that a composite question relating to
soil tests and primary nutrients (N, P, and K) is descriptive. Farmers did not state
exact nutrient levels for healthy and unhealthy soils, rather soil tests served more
to identify the amount of corrective action required to build or maintain soil fer-
Table 3–1. Representative descriptive terms for top 20 soil health properties (from Romig et al., 1995).

Soil health Descriptions

Rank property Healthy Unhealthy
1 Organic mat ter As high as possible, at soil's potential, manure, composit, 3%, 2%, Rough, lack of organic matter, less, low.
7-8%, putting more back.
2 Crop appear ance i
Green, healthy, uniform, lush, dense stand, tall, larger, sturdy, Yellow, stunted corn, small, poor color, poorer, lack of green, light
stout, proper color, darker, good crop. green, streaks in field.
3 Erosion Wouldn't erode, water and wind not taking soil, prevented, stays Blows sooner, washes, topsoil's lost, erodes more, clouds of d ust,
in place, less, slowed down, delayed. ravines, runs bad, any, easier.
4 Earthworms Fishing and red worms present, see after rain, a lot, angle worms, Not there, don't work, can't find, no holes, lack of, killed by i nsecti-
see holes and castings, see d uring plowing. cides or anhydrous, void.
5 Drainage Water goes away, fast, better, nio ponding, moves throug;h, takes a Tight, waterlogged, drains too fast, ponding, no outlet for watisr,
<
lot of rain, drains properly, dries out. won't drain, slope, poor, saturated.
FARMER-BASED ASSESSMENT

6 Tillage ease One pass and ready, breaks up, mellow, easier, smooth, crumbles, Never works down, needs more disking, lumps, slabs, shiney, pulls
flows, plow a gear faster, minimum. hard, worked wet, overworked.
7 Soil structune Won't roll out of hand, crumbl;y, loose, holds together, g;ranular. Hard, doesn't hold together, lumpy, falls apart, massive, clodd
lumpy, clumpy, tight, compacted, powder.
8 pH 7.0, 6.7-6.8, 6.2-6.7, balanced,, neutralize. <6.0, high, nothing works, wrong, too low, high acidity.
9 Soil test Up to recommendations, high, elevated, complete, where it be- Law or minimum at work.
longs, every year or two, stay up with soil test.
10 Yield 150-180 bu corn, 60 bu beans, 30^0% higher, +10 bu ; acre-1, 110 bu corn, 150 bu corn, 35 bu beans, 20-50% less, don't ge t much
better 5 yr average, significantly higher, off, down, reduced, low.
11 Compaction Doesn't pack down, not compa ded, stays loose, not out there Compacted, plow layer, packs down, hardpan, plowsoil, tight, can't
when wet. get into it, packed.
L2 Infiltration Water doesn't stand, absorbs, v/ater moves into soil, soa ks, rapid, Water runs off soil, sits on top, water stands, doesn't absorb, p»ud-
no ponding, fast, spongy. dies, nonporous.
L3 Soil color Dark, black, dark brown, gray,holds dark color. Orange, brown, light, white, red, blue-gray, subsoil color, blea ched,
sandy colored, light brown, pale, anemic, gray.
L4 N Put on less, manure, as requireid, compost, slurry, more ;available, Too much N, chemical N, commercial fertilizers burn ground, anhy-
organic N, organic matter. drous, sludge.
L5 Water retentiion Holds moisture, get by with lesis, retains more, moisture travels, Too much water, doesn't hold water, dries out, too wet or dry,
gives and takes water freely, conserving. droughty, stays wet, runs out of moisture, poor.
L6 P As required.
1 Nutrient defi ciency Has what it needs, no shortage of elements, no spots on leaves, Yellow, purple, discoloration in leaves, lodging, crop falls off, strip-
L7 ping, brown streaks, firings on bottom, blight.
Decompositi on Breaks down, decays, rots in 4--5 mon, manure part of soil in See stalks from last year, doesn't break down, manure plows up next
[8 1 yr, disappears, not too fast., 2/3 gone in year. year.
J K As required.
43

.9 Roots Larger, spread out, grow down,, white, deep, numerous,,good Don't penetrate, undeveloped, balled up, grow crossways discol-
to penetration, full, lots of feeders, branched out. ored, diseased, at hard angles, shallow, short.
44 ROMIG ET AL.

tility (Romig et al., 1995). For the scorecard, the soil test question reflects this
farmer point of view.

Content and Structure

The scorecard is a field tool to assess and monitor soil quality and health.
The scorecard is farmer-based, reflecting the priorities, language, and intent of
the growers we interviewed. It is integrative in its evaluation, scoring soil attrib-
utes, and properties of the systems supported by the soil (plant, animal, and water
systems) that contribute to a farmer's diagnosis of soil health. The scorecard uses
primarily sensory-perceived or descriptive indicator properties rather than results
from laboratory analyses.
Structurally, the scorecard is in the form of a booklet, modeled after
descriptive questionnaires, schedules, and profiles used in human health and
quality-of-life studies (Bowling, 1991; McDowell & Newell, 1987). Following a
cover page, explanatory text reviews the procedures of the scorecard's use. Forty-
three soil health indicator properties are described at the three levels of health and
scored in the remaining pages. Associated with each indicator property is a super-
script number that denotes its relative importance and rank with respect to the
other properties in the scorecard. Users are encouraged to score each indicator
anywhere along the 0 to 4 point scale in intervals preferably no smaller than 0.5
to reflect more accurately the observed condition of the soil being graded.
The scorecard integrates observations made throughout the growing season
and is best completed near or just following harvest. Attributes that are expressed
either seasonally or periodically (e.g., soil smell, seed germination, or infiltration)
should be recorded when observed to increase the precision of the instrument.
On the final page is a guide for users to interpret the results of the score-
card. First distributions of soil health indicator properties within the three health
categories – healthy (3 to 4), impaired (1.5 to 2.5), and unhealthy (0 to 1) – are
tallied and percentages are calculated. Ideally, most, if not all, indicators should
lie within the healthy category. Even if a high percentage of indicators were grad-
ed as healthy, it is important to note those properties that scored low and may be
serious problems requiring attention. Those indicators that were scored as
unhealthy probably need immediate corrective action. For indicators either in the
unhealthy or impaired category, it is necessary to consider what caused the con-
dition in the first place and how it might be attenuated. For indicators that were
assessed as impaired, the users should judge from a historical perspective,
whether that particular property's condition is deteriorating or improving. Users
also may wish to focus more on those properties farmers considered most impor-
tant as indicated by its relative rank in superscript.
Final scores could be calculated to assess the entire farming system or sep-
arate scores for each target system could be computed for comparative analysis
between the scores for the soil system and the plant, animal, and water systems.
This approach would provide a gross estimate of management effects on soil
health over time. Such simplified procedures, however, may obscure certain
properties that were graded low and may need attention. Furthermore, the com-
putation of an overall final score is built on the assumption that all indicators are
FARMER-BASED ASSESSMENT 45

equally important to a soil health assessment. The scorecard in its present form
does not recognize the relative importance properties have with respect to one
another and weight the score of each indicator accordingly. The benefit of differ-
ential weighting schemes in indices similar to the scorecard has been a topic of
discussion in the health and behavioral sciences for some time and with little
resolve (Wang & Stanley, 1970; Wainer, 1976; Skinner & Lei, 1980; Streiner &
Norman, 1989). The use of differentially weighting indicators in the scorecard
will only become feasible as measurement models of complex soil processes
related to soil quality become more fully developed.

Pretesting

Initial field tests in cropping system comparison trials have shown that the
information generated by the scorecard holds promise. The scorecard identified
several descriptive indicators in both soil and plant systems that demonstrated
variable responses under different management regimes, even on soils considered
very resilient (Garlynd et al., 1995). These indicators included earthworms, soil
structure, decomposition, surface crusting, soil smell, aeration, crop appearance,
yield (especially when the growing season was stressful), nutrient deficiencies,
mature crop, plant stems, and plant resistance to disease and pests. Distributions
of indicator properties within the health categories have exhibited a greater per-
centage of healthy indicators in low-input cropping systems as compared with
continuous corn (Zea mays L.). For the top 12 indicator properties (Romig et al.,
1995), soil health scores for continuous corn and lower-input cropping strategies
were uniformly high, except for earthworms (Lumbricus terrestris) that scored
lower in continuous corn treatments (Garlynd et al., 1995).
Pretests of the scorecard with farmers showed that farmers may be inher-
ently biased when grading their own soils. Besides the affection they have for
their farm, farmer bias may be due to the integrative nature of their knowledge.
Farmer knowledge of soils is often intimately connected to management prac-
tices, both past and present. When judging any property they also may be judg-
ing the practices carried out on the field (Romig et al., 1995). Users need to be
aware that without maintaining some level of objectivity, the effectiveness of the
scorecard will be compromised. Furthermore, farmers are most familiar with the
soils in their locality and may have limited experience with a number of different
soil types. Full potential of the scorecard will be realized as users become famil-
iar with the natural variability of soil descriptive indicators for the region where
the scorecard was developed.
Initial reaction of farmers to the scorecard has been generally positive. One
of the shortcomings of the scorecard, however, is that it is driven solely by the
results of these 28 interviews, and may contain biases with respect to soils, farm
type, and input strategy. Organic fruit and vegetable producers, for example, have
commented on the scorecard's conventional audience and its focus on cow crops.
Therefore, caution is advised when applying the scorecard to farming systems in
locations unlike those from which it was developed. Modification of the score-
card for other cropping systems and other regions would require structured input
from additional farmers.
46 ROMIG ET AL.

Applications

The scorecard provides a farmer-based assessment of soil health that has

inherent value as a field tool for farmers, extension agents, crop consultants, and
scientists. The scorecard is flexible not only for these different user groups, but
for different farming systems as well, because only the indicators relevant to the
farming system need to be evaluated.
The scorecard helps farmers collect and assess information that is both per-
tinent to their operation and at their fingertips. Moreover, the scorecard provides
the farmer with an integrative assessment of the overall effect of management on
soil health, helps monitor soil health over time, and identifies the individual prop-
erties that are being impacted either negatively or positively by soil management.
This, in turn, can help farmers conduct their own analyses of a soil's condition
and make adjustments in management if necessary.
In a research setting comparing management strategies, the scorecard pro-
vides an initial appraisal of the effects of different cropping systems on individ-
ual properties as well as a more integrated understanding of a soil's health under
different systems. The farmer knowledge embodied in the scorecard provides
additional insight that may help ground scientific interpretations, predictions, and
models of the effects of management on specific properties and functions of soil
quality. The scorecard also can be used to identify descriptive indicators, and
their corresponding analytical properties, that may be sensitive to changes in
management. For example, preliminary use of the scorecard in trials monitoring
the transition from continuous corn to more complex rotations indicates that the
number of earthworms, their castings, and holes increase in more varied rotations
(Garlynd et al., 1995). Another avenue researchers may wish to examine is the
comparison of the scores of soil health indicators relevant to a given soil quality
function (i.e., water relationships, crop performance, and nutrient cycling), per-
haps in a framework similar to that presented in Karlen et al. (1994). By examin-
ing indicator properties that contribute to a given function, it may be possible to
better articulate the relationship between certain management practices and soil
functions. The scorecard has additional merit as a reference base for soil quality
assessments and may assist with the interpretation of quantitative data by helping
calibrate data from laboratory analyses to field conditions.
The scorecard has value in that it provides linkages between scientist and
farmer assessment of soil health and quality. First, the integrative nature of the
scorecard, assessing soil, plants, animals or humans, and water indicator proper-
ties, has conceptual value for scientists, particularly in the development and val-
idation of indices addressing the functions of soil quality (Harris & Bezdicek,
1994; Romig et al., 1995). Moreover, the scorecard may help acclimate the sci-
entific community to soil health and quality indices that qualitatively integrate
analytical and/or descriptive data. Second, despite recognition of the need to
move beyond disciplinary boundaries and pursue holistic, accessible indicators
(Doran & Parkin, 1994), nonsoil and descriptive indicator properties for soil qual-
ity characterization have received little attention. In this regard, the scorecard
provides a base to identify and develop both descriptive and analytical soil and
nonsoil indicators for soil health and quality assessment. Furthermore, farmer use
FARMER-BASED ASSESSMENT 47

of descriptive soil health indicator properties supports the inclusion of corre-

sponding analytical properties in quantitative soil quality indices. Third, the
scorecard provides a source of descriptive data for integrated soil health and qual-
ity scorecards (Harris et al., 1996).
It is important to recognize that the Scorecard has limitations that were
inherited from the interview data from which it was developed. One shortcoming
is that the scorecard represents the unchallenged perceptions of a relatively small
group of farmers. While confident that the scorecard addresses many of the cru-
cial indicator properties of soil health and quality, continued dialogue between
farmers and scientists is necessary to advance our collective understanding and
will undoubtedly result in modifications of the structure and composition of the
scorecard, particularly in the scope and nature of the plant, animal, and water
indicators of soil quality.
A second limitation to the scorecard may be its narrow frame of reference,
especially in light of the expanded definition of soil quality. Given that it reflects
the nature of soil health that farmers perceive as important, the scorecard is inher-
ently focused upon crop production and may fail to recognize critical aspects of
soil quality with respect to environmental protection. This is evident in that only
a small minority of soil health indicators identified by the farmers are directly
associated with environmental quality, specifically erosion, surface water appear-
ance and chemicals in groundwater. Similarly, a soil test with high nutrient levels
for crop production would be graded high by the scorecard, while that same soil
test might receive a much lower rating with regards to water quality.
Despite these concerns, the strength and utility of the Wisconsin Soil Health
Scorecard should be emphasized. The scorecard provides a mechanism to assess
and monitor individual descriptive soil quality indicators in the field, acts as a
communication bridge between farmers and scientists, and gives valuable
insights into meaningful analytical soil quality indicators.

REFERENCES
Acton, D.F., and G.A. Padbury. 1993. Introduction. p. 2.1–2.10. In D.F. Acton (ed.) A program to
assess and monitor soil quality in Canada. Res. Branch, Agric. Canada. CLBRR No. 93–49.
Agric. Canada, Ottawa, Ontario, Canada.
Arshad, M.A., and G.M. Coen. 1992. Characterization of soil quality: Physical and chemical criteria.
Am. J. Altern. Agric. 7:25–30.
Blalock, H.M. 1972. Social statistics. McGraw-Hill, NewYork.
Bowling, A. 1991. Measuring health: A review of quality of life measurement scales. Open Univ.
Press, Philadelphia, PA.
Brown, T.C., and T.C. Daniel. 1990. Scaling of ratings: Concepts and methods. U.S. Dep. of Agric.
Forest Serv. Res. Pap. RM-293. Rocky Mountain For. and Range Exp. Stn., Fort Collins, CO.
Chambers, R. 1993. Methods of analysis by farmers. J. Farming Syst. Res. 4:87–101.
Culyer, A.J., 1983. Health indicators. St. Martin's Press, New York.
Cunningham, L., J. Doll, J. Hall, D. Mueller, J. Posner, R. Saxby, and A. Wood. 1992. The Wiscon-
sin integrated cropping system trial first report. Univ. of Wisconsin, Arlington.
Doran, J.W., and T.B. Parkin. 1994. Defining and assessing soil quality. p. 3–21. In J.W. Doran et al.
(ed.). Defining soil quality for a sustainable environment. SSSA Spec. Publ. 35. SSSA, Madi-
son, WI.
Garlynd, M.J., A.V. Karakov, D.E. Romig, and R.F. Harris. 1994. Descriptive and analytical charac-
terization of soil quality/health. p. 159–168. In J.W. Doran et al. (ed.) Defining soil quality for
a sustainable environment. SSSA Spec. Publ. 35. SSSA, Madison, WI.
 48 ROMIG ET AL.

Garlynd, M.J., D.E. Romig, and R.F. Harris. 1995. Effect of a cropping systems shift from continu-
ous corn on descriptive and analytical indicators of soil quality. p. 58 In Agronomy abstracts.
ASA, Madison, WI
Granatstein, D., and D.F. Bezdicek. 1992. The need for a soil quality index: Local and regional per-
spectives. Am. J. Altern. Agric. 7:12–16.
Grossman, R.B., and F.B. Pringle. 1987. Describing surface soil properties – their seasonal changes
and interpretations for management. p. 57–75. In Soil survey techniques. SSSA Spec. Publ.
35. SSSA, Madison, WI.
Haberern, J. 1992. A soil health index. J. Soil Water Conserv. 47:6.
Harris, R.E, and D.F. Bezdicek. 1994. Descriptive aspects of soil quality/health. p. 23–35. In J.W.
Doran et al. (ed.) Defining soil quality for a sustainable environment. SSSA Spec. Publ. 35.
SSSA, Madison, WI.
Harris, R.F., D.L. Karlen, and D.J. Mulla. 1996. A conceptual framework for assessment and man-
agement of soil quality and health. p. 61–82. In J.W. Doran and A.J Jones (ed.) Methods for
assessment of soil quality and health. SSSA Spec. Publ. 49. SSSA, Madison, WI.
Karlen, D., and D. Stott. 1994. A framework for evaluating physical and chemical indicators of soil
quality. p. 53–72. In J.W. Doran et al. (ed.) Defining soil quality for a sustainable environment.
SSSA Spec. Publ. 35. SSSA, Madison, WI.
Karlen, D.L., M.J. Mausbach, J.W. Doran, R.G. Cline, R.F. Harris, and G.E. Schuman. 1997. Soil
quality: Concept, rationale, and research needs. Soil Sci. Soc. Am. J. 60:4–10.
Karlen, D.L., N.C. Wollenhaupt, D.C. Erbach, E.C. Berry, J.B. Swan, N.S. Eash, and J.L. Jordahl.
1994. Crop residue effects on soil quality following 10 years of no-till corn. Soil Tillage Res.
31:149–167.
Kelling, K.A. (ed.). 1989. Proc. of the University of Wisconsin sustainable agriculture "Listening"
meetings. Wisconsin Agric. Exp. Stn. and Wisconsin Coop. Ext. Serv., Madison.
Larson, J.S., 1991. The measurement of health: Concepts and indicators. Greenwood Press, New
York.
Larson, W.E., and F.J. Pierce. 1991. Conservation and enhancement of soil quality. In Evaluation for
sustainable land management in the developing world. p. 175–203. In Proc. of the Int. Work-
shop on Evaluation for Sustainable Land Management in the Developing World, Chiang Rai.
15–21 Sept. 1991. Int. Board of Soil Res. and Manage., Bangkok, Thailand.
McDowell, I., and C. Newell. 1987. Measuring health: A guide to rating scales and questionnaires.
Oxford Univ. Press, New York.
Parr, J.F., R.I. Papendick, S.B. Hornick, and R.E. Meyer. 1992. Soil quality: Attributes and relation-
ship to alternative and sustainable agriculture. Am. J. Altern. Agric. 7:5–11.
Reganold, J.P., A.S. Palmer, J.C. Lockhart, and A.N. Macgregor. 1993. Soil quality and financial per-
formance of biodynamic and conventional farms in New Zealand. Science (Washington, DC)
260:344–349.
Romig, D.E. 1995. Farmer knowledge of soil health and its role in soil quality assessment. M.S. the-
sis. Univ. of Wisconsin, Madison.
Romig, D.E., M.J. Garlynd, R.F. Harris, and K. McSweeney. 1995. How farmers assess soil health
and quality. J. Soil Water Conserv. 50:225–232.
Rosser, R. 1983. Issues of measurement in the design of health indicators: A review. p. 34–43. In A.J.
Culyer (ed.) Health indicators. St. Martin's Press, New York.
Sharman, A.K. 1989. Qualitative and quantitative land evaluation for rainfed maize in subhumid trop-
ical and subtropical climates. Adv. Soil Sci. 9:147–156.
Skinner, H.A., and H. Lei. 1980. Differential weights in life change research: Useful or irrelevant.
Psychosomatic Medicine 42:367–370.
Stevens, S.S. 1960. Ratio scales, partition scales, and confusion scales. p. 49–66. In H. Gulliksen and
S. Merrick (ed.) Psychological scaling: Theory and applications. John Wiley & Sons, New
York.
Stevens, S.S. 1975. Psychophysics. John Wiley & Sons, New York.
Streiner, D.L., and G.R. Norman, 1989. Health measurement scales: A practical guide to their devel-
opment and use. Oxford Univ. Press, New York.
Thrupp, L.A. 1989. Legitimizing local knowledge: From displacement to empowerment for third
world people. Agric. Human Values 6(3): 13–24.
Torgerson, W.S. 1958. Theory and methods of scaling. John Wiley & Sons, New York.
Wainer, H. 1976. Estimating coefficients in linear models: It don't make no nevermind. Psychol. Bull.
83:213–217.
Wang, M.W., and J.C. Stanley. 1970. Differential weighting: A review of methods and empirical stud-
ies. Rev. Educ. Res. 40:663–705.
FARMER-BASED ASSESSMENT 49
50 ROMIG ET AL.