animal

Animal (2008), 2:8, pp 12031214 & The Animal Consortium 2008

doi:10.1017/S1751731108002280

Meta-analyses of experimental data in animal nutrition*

D. Sauvant1-, P. Schmidely1, J. J. Daudin2 and N. R. St-Pierre3
1
UMR Physiologie de la Nutrition et Alimentation, INRA-AgroParisTech, 16 rue Claude Bernard, 75231 Paris Cedex 05, France; 2ENGREF, Mathematique et
Informatique Appliquee, INRA-AgroParisTech, 16 rue Claude Bernard, 75231 Paris Cedex 05, France; 3Department of Animal Sciences, The Ohio State University,
2029 Fyffe Rd., Columbus, OH-43210, USA

(Received 6 April 2007; Accepted 14 March 2008)

Research in animal sciences, especially nutrition, increasingly requires processing and modeling of databases. In certain areas
of research, the number of publications and results per publications is increasing, thus periodically requiring quantitative
summarizations of literature data. In such instances, statistical methods dealing with the analysis of summary (literature) data,
known as meta-analyses, must be used. The implementation of a meta-analysis is done in several phases. The rst phase
concerns the denition of the study objectives and the identication of the criteria to be used in the selection of prior
publications to be used in the construction of the database. Publications must be scrupulously evaluated before being entered
into the database. During this phase, it is important to carefully encode each record with pertinent descriptive attributes
(experiments, treatments, etc.) to serve as important reference points for the rest of the analysis. Databases from literature
data are inherently unbalanced statistically, leading to considerable analytical and interpretation difculties; missing data are
frequent, and data structures are not the outcomes of a classical experimental system. An initial graphical examination of the
data is recommended to enhance a global view as well as to identify specic relationships to be investigated. This phase is
followed by a study of the meta-system made up of the database to be interpreted. These steps condition the denition of the
applied statistical model. Variance decomposition must account for inter- and intrastudy sources; dependent and independent
variables must be identied either as discrete (qualitative) or continuous (quantitative). Effects must be dened as either xed
or random. Often, observations must be weighed to account for differences in the precision of the reported means. Once model
parameters are estimated, extensive analyses of residual variations must be performed. The roles of the different treatments
and studies in the results obtained must be identied. Often, this requires returning to an earlier step in the process. Thus,
meta-analyses have inherent heuristic qualities.
Keywords: meta-analysis, mixed models, nutrition

Introduction
The research environment in the animal sciences, especially
nutrition, has markedly evolved in the recent past. In
particular, there is a noticeable increase in the number
of publications, each containing an increasing number of
quantitative measurements. Meanwhile, treatments often
have smaller effects on the systems being studied than in
the past. Additionally, controlled and non-controlled factors,
such as the basal plane of nutrition, vary from study
to study, thus requiring at some point a quantitative summarization of past research.
*Salaries and research support were provided by state and federal funds
appropriated to the Ohio Agricultural Research and Development Center, the
Ohio State University.
E-mail: sauvant@agroparistech.fr

Fundamental research in the basic animal science disciplines generates results that increasingly are at a much
lower level of aggregation than those of applied research
(organs, whole animals), thus supporting the necessity of
integrative research. Research stakeholders, those who ultimately use the research outcome, increasingly want more
quantitative knowledge, particularly on animal response
to diet, and of better precision. Forecasting and decisionsupport software require quantitative information. Additionally, research prioritization by funding sources may force
abandoning active research activities in certain elds. In such
instances, meta-analyses can still support discovery activities
based on the published literature.
The objectives of this paper are to describe the application
of meta-analytic methods to animal nutrition studies, including the development and validation of literature-derived
1203

Sauvant, Schmidely, Daudin and St-Pierre

databases, and the quantitative techniques used to extract
the quantitative information.

>

Denitions and nature of problems

Limits to classical approaches

Results from a single classical experiment cannot be the
basis for a large inference space because the conditions
under which observations are made in a single experiment
are forcibly very narrow, i.e. specic to the study in
question. Such studies are ideal to demonstrate cause and
effect, to test specic hypothesis regarding mechanisms
and modes of action. In essence, a single experimentation
measures the effects of one or a very few factors while
maintaining all other factors as constant as possible. Often,
experiments are repeated by others to verify the generality
and repeatability of the observations that were made,
as well as to challenge the range of applicability of the
observed results and conclusions. Hence, it is not uncommon that over time, tens of studies are published even on a
relatively narrow subject. In this context, there is a need to
summarize the ndings across all the published studies.
Meta-analytic methods are concerned with how best to
achieve this integration process.
The classical approach to synthesizing scientic knowledge has predominantly been based on qualitative literature reviews. A limitation of this approach is the obvious
subjectivity involved in the process. The authors subjectively
weigh outcomes from different studies. Criteria for the
inclusion or non-inclusion of studies are ill dened at best.
Different authors can draw dramatically different conclusions from the same initial set of published studies. Additionally, the limitation of the human brain to differentiate
the effects of many factors becomes very apparent once the
number of publications involved exceeds 12 to 15 studies.
Denitions and objectives of meta-analyses
Meta-analyses use objective, scientic methods based on
statistics to summarize and quantify knowledge acquired
through prior published research. Meta-analytic methods
were initially developed in psychology, medicine and social
sciences a few decades ago. Meta-analytic reviews have
been more recent and much less frequent in nutrition.
In general, meta-analyses are conducted for one of the
following four objectives:
>

For global hypothesis testing, such as testing for the

effect of a certain drug or of a feed additive using the
outcomes of several publications that had as an objective the testing of such an effect. This was by far the
predominant objective of the rst meta-analyses published (Mantel and Haenszel, 1959; Glass, 1976). Early
on, it was realized that many studies lacked statistical
power for statistical testing, so that the aggregation of
results from many studies would lead to much greater
power (hence lower type II error), more precise point
estimation of the magnitude of effects and narrower
condence intervals of the estimated effects.

1204

>

>

For empirical modeling of biological responses, such as the

response of animals to nutritional practices. Because the
data extracted from many publications cover a much wider
set of experimental conditions than those of each individual
study, conclusions and models derived from the whole
set have a much greater likelihood of yielding relevant
predictions to assist decision-makers. These meta-analytic
activities have led to a new paradigm suggested by
Sauvant (1992 and 1994) as a law of multiple responses to
changes in feeding practices. For example, alteration in
feeding practices impact feed efciency, product quality,
the environment, animal welfare, etc., and signicant
progress can be made in this type of holistic research using
meta-analyses. There are numerous examples of such
applications of meta-analytical methods in recent nutrition
publications, such as the quantication of the physiological
response of ruminants to types of dietary starch (Offner and
Sauvant, 2004), to the supply of dietary N (Rico-Gomez
and Faverdin, 2001), to dietary fats (Schmidely et al., 2008)
and rumen defaunation (Euge`ne et al., 2004). Others
have used meta-analyses to quantify phosphorus ux in
ruminants (Bravo et al., 2003) or carcass characteristics to
various factors (McPhee et al., 2006).
For collective summarizations of measurements that
only had a secondary or minor role in prior experiments.
Frequently, results are reported in publications with the
objective of supporting the hypothesis or observations
related to the effect of one or a few experimental factors.
For example, ruminal volatile fatty acid (VFA) concentrations are reported in studies investigating the effects of
dietary starch, or forage types. None of these studies
have as an objective the prediction of ruminal VFAs. But
the aggregation of measurements from many studies
can lead to a better understanding of factors controlling
VFA concentrations, or allow the establishment of new
research hypotheses.
In mechanistic modeling, for parameter estimates and
estimates of initial conditions of state variables. Mechanistic
models require parameterization, and meta-analyses offer a
mechanism of estimation that makes parameter estimation
more precise and more applicable to a broader range of
conditions. Meta-analyses can also be used for external
model evaluation (Lovatto and Sauvant, 2002; Sauvant and
Martin, 2004), or for a critical comparison of alternate
mechanistic models (Offner and Sauvant, 2004; Offner
et al., 2003).

Types of data and factors in meta-analyses

As in conventional statistical analyses, dependent variables in
meta-analyses can be of various types such as binary [0, 1]
(e.g. for pregnancy), counts or percentages, categoricalordinal (good, very good, excellent), and continuous, which is
the most frequent type in meta-analyses related to nutrition.
Independent factors (or variables) may be modeled using
a xed or random effect. In general, factors related to
nutrition (grain types, dry matter intake (DMI), etc.) should
be considered as xed effects factors. The study effect can

Meta-analyses in nutrition
be considered as either random or xed. If a dataset
comprised many individual studies from multiple research
centers, the study effect should be considered random
because each study is conceptually a random outcome from
a large population of studies to which inference is to be
made (St-Pierre, 2001). This is especially important if the
meta-analysis has for objective the empirical modeling of
biological responses, or the collective summarizations of
measurements that only had a secondary or minor role in
prior experiments, because it is likely that the researcher in
those instances has a targeted range of inference much
larger than the limited conditions represented by the specic studies. There are instances, however, where each
experiment can be considered as an outcome each from a
different population. In such instances, the levels of study
or trial are, in essence, considered arbitrarily chosen by
the research community, and the study effect must then be
considered xed. In such an instance, the range of inference
for the meta-analysis is limited to the domain of the specic
experiments in the dataset. This is of little concern if the
objective of the meta-analysis is that of global hypothesis
testing, but it does severely limit the applicability of its
results for other objectives.

Difculties inherent to the data

The meta-analytic database is best conceptualized with
rows representing treatments, groups or lots, while the
columns consist of the measured variables (those for which
least-squares means are reported) and characteristics (class
levels) of the treatments or trials. A primary characteristic
of most meta-analytic databases is the large frequency
of missing data in the table. This reduces the possibility of
using large multi-dimensional descriptive models, and
generally forces the adoption of models with a small subset
of independent variables, frequently two. Additionally, the
design of the meta-analytic data, sometimes referred to as
the meta-design, is not determined prior to the data collection as in classic randomized experiments. Consequently,
meta-analytic data are generally severally unbalanced and
factor effects are far from being orthogonal (independent).
This leads to unique statistical estimation problems similar
to those observed in observational studies, such as leverage
points, near collinearity, and even complete factor disconnectedness, thus prohibiting the testing of the effects that
are completely confounded with others.
Table 1 shows an example of factor disconnectedness,
where two factors, each taking three possible levels, are
being investigated. In this example, factors A and B are
disconnected because one cannot join all bordering pairs of
cells with both horizontal and vertical links. Consequently,
even in a model without any interaction terms, the effect of
the third level of A cannot be estimated separately from the
effect of the third level of B. This would be diagnosed
differently depending on the software used with different
combination of error messages in the log, zero degrees of
freedom for some effects in the ANOVA table, or a missing
value for the statistic used for testing.

Table 1 Example of factor disconnectedness

Factor A
Factor B
1
2
3

x
x

x
x

In general, the variance between studies is large compared to the variance within studies, hence underlying the
importance of including the study effect into the metaanalytical model. The study effect represents the sum of the
effects of many factors that differ between studies, but
factors that are not in the model because they either were
not measured, or have been excluded from the model, or
for which the functional form in the model is inadequately
representing the true but unknown functional form (e.g. the
model assumes a linear relationship between the dependent and one independent continuous variable whereas the
true relationship is nonlinear). In the absence of interactions
between design variables (e.g. studies) and the covariates
(e.g. all model variables of interest), parameter estimates
for the covariates are unbiased, but the study effect can add
a large uncertainty to future predictions (St-Pierre, 2001).
The presence of signicant interactions between studies
and at least one covariate is more problematic since this
indicates that the effect of the covariate is dependent on
the study, implying that the effect of a factor is dependent
on the levels of unidentied factors.

Steps in the meta-analytic process

There are several inherent steps to meta-analyses, the
important ones are summarized in Figure 1. An important
aspect of this type of analyses is the iteration process,
which is under the control of the analyst. This circular
pattern where prior steps are re-visited and rened is an
important aspect of meta-analyses and contributes much to
their heuristic characteristic.
Objectives of the study
Establishing a clear set of study objectives is a critical step
that guides most ulterior decisions such as the database
structure, data ltering, weighing of observations and
choice of the statistical model. Objectives can cover a wide
range of targets, ranging from preliminary analyses to
identify potential factors affecting a system, thus serving an
important role to the formulation of research hypotheses in
future experiments, to the quantication of the effect of a
nutritional factor such as a specic feed additive.
Data entry
Results from prior research found in the literature must be
entered in a database. The structure and coding of the
database must include numerous variables identifying
the experimental objectives of all experiments selected.
Hence, numerous columns are potentially added to code
1205

Sauvant, Schmidely, Daudin and St-Pierre

OBJECTIVES

CONCEPTUAL
Basis

oo

SELECTION
OF EXPERIMENTS
& DATA ENTRY
GRAPHICAL
ANALYSIS

Literature search,

Experiments
Survey,

DETERMINATION OF
META-DESIGN

SELECTION OF
STATISTICAL
MODEL

Adjustments
USERS
POST-ANALYSIS
EVALUATION

Figure 1 Schematic representation of the meta-analytic process.

each objective found in all the sources of experimental

data. This coding is necessary so as to avoid the improper
aggregation of results across studies with very different
objectives. During this coding phase, the analyst may chose
to transform a continuous variable to a discrete variable
with n levels coded in a single column with levels of the
discrete variable as entries, or in n columns with 0 to 1
entries to be used as dummy variables in the meta-analytic
model. Different criteria can guide the selection of classes,
such as equidistant classes, or classes with equal frequencies or probability of occurrence. The important point is
that the sum of these descriptive columns must entirely
characterize the objectives of all studies used.

Data ltering
There are at least three steps necessary to effective
data ltering. First, the analyst must ensure that the study
under consideration is coherent with the objectives of the
meta-analysis. That is, the meta-analytic objectives dictate
that some traits must be measured and reported. If, for
example, the objective of meta-analysis is to quantify the
relationship between dietary neutral detergent ber (NDF)
concentration and DMI, then one must ensure that both
NDF concentration and DMI were measured and reported
in all studies. The second step consists of a thorough and
critical review of each publication under consideration,
focusing on the detection of errors in the reporting of
quantitative results. This underlines the importance of
having a highly trained professional involved in this phase
of the study. Only after publications have passed this
expert quality lter should their results be entered in the
database. Verication of data entries is then another
essential component to the process. In this third step, it is
important to ensure that a selected publication does not
appear to be an outlier with respect to the characteristics
and relations under consideration.
1206

Preliminary graphical analyses

A thorough visual analysis of the data is an essential step to
the meta-analytic process. During this phase, the analyst
can form a global view regarding the coherence and heterogeneity of the data, as well as to the nature and relative
importance of the inter- and intrastudy relationships of
prospective variables taken two at a time. Systematic graphical analyses should lead to specic hypotheses and to
the initial selection of alternate statistical models. Graphics
can also help in identifying observations that appear unique
or even outliers to the mass of all other observations. The
general structure of relationships can also be identied,
such as linear v. nonlinear relationships as well as the
presence of interactions. As an example, Figure 2 shows a
ctitious example of an intrastudy curvilinear relationship
between two variables in the presence of a negative
interstudy effect. In this example, the negative interstudy
effect associated with the X variable indicates the presence
of a latent variable that differed across studies, and that
interacts with the X variable. A review of the specic
characteristics of studies (3) and (2) in this example might
help to identify this latent variable. This visualization
phase of the data should always be taken as a preliminary
step to the statistical analysis and not as conclusive evidence. The reason is that as the multi-dimensions of the
data are collapsed into two- or possibly three-dimensional
graphics, the unbalance that clearly is an inherent characteristic of meta-analytic data can lead to false visual
relationships. This is because XY graphics do not correct
the observations for the effects of all other variables that
can affect Y.
Graphical analyses should also be done with regard to
the joint coverage of predictor variables, identifying their
possible ranges, plausible ranges and joint distributions, all
being closely related to the inference range. Similar XY
graphics should be drawn to explore the relationships

Meta-analyses in nutrition
1100

Chewing (Min/d)

1000
900
800
700
600
500
400
300
20

30

40

50

60

70

NDF (% of DM)

between predictor variables taken two at a time. In such

graphics, the presence of any linear trends indicates
correlations between predictor variables. Strong positive
or negative correlations of predictor variables have two
undesirable effects. First, they may induce near collinearity,
implying that the effect of one predictor cannot be uniquely
identied (i.e. is nearly confounded with the effect of
another predictor). In such instances, the statistical model
can include only one of the two predictors at a time.
Second, the range of a predictor X1 given a level of a
second predictor X2 is considerably less than the unconditional range of predictor X1. In these instances, although
the range of a predictor appears considerable in a univariate setting, its effective range is actually very much
reduced in the multivariate space.
Figure 3 illustrates some of these concepts using an
actual set of data on chewing activity in cattle and the NDF
content of the diet. Visually, one concludes that both the
intra- and interstudy relationships between chewing activity
and diet NDF are nonlinear. This observation can be formally
tested using the statistical methods to be outlined later.
In this example, the possible NDF range is 0% to 100%,
whereas its plausible range is more likely between 35% and
60%. Interestingly, the gure illustrates that experimental
measurements were frequently in the 20% to 40% and
45% to 55% ranges, leaving a hole with very few observations in the 40% to 45% range.

Study of the experimental meta-design

The meta-design is determined by the structure of the
experiments for each of the predictor variables. To characterize the meta-design, numerous steps must take
place before and after the statistical analyses. The specic
steps depend on the number of predictor variables in the
model.
One predictor variable
> The experimental design used in each of the studies
forming the database must be identied and coded,

Figure 3 Effect of dietary neutral detergent ber (NDF) content on

chewing activity in cattle. Data are from published experiments where the
NDF content of the diet was the experimental treatment.

60
n = 517
m = 38.4
s = 13

50
Frequency

Figure 2 Example of a curvilinear intrastudy effect in the presence of a

linear negative interstudy effect.

40
30
20
10
0
10

20

30

40

50

60

70

80

NDF (% of DM)

Figure 4 Frequency plot of neutral detergent ber (NDF) content of diets

across experimental treatments.

>

>

and their relative frequencies calculated. This information

can be valuable during the interpretation of the results.
Frequency plots (histograms) of the predictor variable can
identify areas of focus of prior research. For example,
Figure 4 shows the frequency distributions of NDF for the
517 treatment groups for the meta-analytic dataset used
to draw Figure 3. Figure 4 indicates a substantial research
effort towards diets containing 30% to 35% NDF, an area
of dietary NDF density that borders the lower limit of
recommended dietary NDF for lactating dairy cows.
Because of the high frequency of observations in the 30%
to 35% NDF range, the a priori expectation is that the
effect of dietary NDF will be estimated most precisely in
this range.
One should also consider the intra- and interstudy
variances for the predictor variable. Small intrastudy
variances reduce the ability of assessing the structural
form of the intrastudy relationship between the predictor
and the dependent variable. Large intrastudy variances,
but with only two levels of the predictor variable, can
hide any potential curvilinear relationships. Figure 5
1207

Sauvant, Schmidely, Daudin and St-Pierre

Intra-experiment root mean square error (%DM)
20

>

10

0
15

25

number of treatments >2

number of treatments = 2

35
45
55
Mean NDF (%DM)

65

75

hi 1=n Xi  Xm 2 =SXi  Xm 2 ;

>

1208

where hi is the leverage value, Xi is the value of the ith

predictor variable and Xm is the mean of all Xi.
Equation (1) clearly indicates that the leverage of an
observation, i.e. its weight in the determination of the
slope, grows with its distance from the mean of the predictor variable. The extension of the leverage point
calculations to more than one predictor variables is
straightforward (St-Pierre and Glamocic, 2000).

Figure 5 Intrastudy root mean square error of dietary neutral detergent

ber (NDF) and mean NDF of experiments designed to study the effects of
NDF in cattle.

shows the intrastudy root mean square error (S) of NDF

as a function of the mean dietary NDF for experiments
with two treatments, and for those with more than
two treatments. The analyst can determine a minimum
threshold of S to exclude experiments with little intrastudy
variation in the predictor variable. In instances where the
study effect is considered random, this is not necessary. In
such instances, the inclusion of studies with little variation
in the predictor variable does little in determining the
relationship between the predictor and the dependent
variable, but adds observations and degrees of freedom to
estimate the variance component associated with the study
effect. When looking at a possible curvilinear intrastudy
relationship, it is intuitive to retain only those experiments
with three or more levels of the predictor variables in the
dataset. However, in Figure 5, this eliminates 40% of all the
observation in the study. In instances where the curvilinear
relationship is only apparent interstudy, one may chose to
retain only experiments with three or more treatments, or
work on the intrastudy deviation between the control and
treatment groups, as suggested by Rico-Gomez and
Faverdin (2001).
Another important aspect at this stage of the analysis is
to determine the signicance of the study effect on the
predictor variables. As explained previously, one must be
very careful regarding the statistical model used to
investigate the intrastudy effect when there is not a good
agreement between the intra- and interstudy effects. In
these instances, the relationship between the predictor
variable and the dependent variable depends on the
study, which itself represents the sums of a great many
factors such as measurement errors, systematic differences in the methods of measurements of the dependent
variable across studies, and, more importantly, latent
variables (hidden) not balanced across experiments. In
those instances, the analyst must exert great caution in

the interpretation of the results, especially regarding the

applicability of these results.
At this stage, even before statistical models are to be
tted and effects tested, it is generally useful to calculate
the leverage of each observation (Tomassone et al.,
1983). Traditionally, in statistics, leverage values are
calculated after the model is tted to the data, but
nothing prohibits the calculation of leverage values at an
earlier stage because their calculations depend only on
the design of the predictor variable in the model. For
example, in the case of the simple linear regression with
n observations, the leverage point for the ith observation
is calculated as

>

In a nal step, the analyst must graphically investigate

the functional form of the relationship between the
dependent variable and the predictor variable.

Two or more predictor variables: In the case of two or

more predictor variables, the analyst must examine graphically and then statistically the inter- and intrastudy
relationships between the predictor variables. Leverage
values should be examined. With xed models (all effects in
the models are xed with the exception of the error term),
variance ination factors (VIFs) should be calculated for
each predictor variable (St-Pierre and Glamocic, 2000).
An equivalent statistic has not been proposed for mixed
models (e.g. when the study effect is random), but
asymptotic theory would support the calculation of the VIFs
for the xed effect factors in cases where the total number
of observations is large. The objective in this phase is
to assess the degree of inter-dependence between the
predictor variables. Because predictor variables in metaanalyses are never structured prior to their determination,
they are always non-orthogonal and, hence, show variable
degrees of inter-dependency. Collinearity determinations
(VIFs) assess ones ability to separate the effects of interdependent factors based on a given set of data. Collinearity
is not model driven, but completely data driven.
Weighing of observations
Because meta-analytic data are extracted from the results
of many experiments conducted under many different
statistical designs and number of experimental units,
the observations (treatment means) have a wide range of
standard errors. Intuition and classical statistical theory
would indicate that observations should be subjected to

Meta-analyses in nutrition
some sort of weighing scheme. Systems used for weighing
observations form two broad categories.

Weighing based on classical statistical theory

Under a general linear model where observations have
heterogeneous but known variances, maximum likelihood
parameters estimates are obtained by weighing each
observation by the inverse of its variance. In the context
of a meta-analysis where observations are least-squares
(or population marginal) means, observations should be
weighed by the inverse of the squares of their standard
errors, which are the standard errors of each mean (s.e.).
Unfortunately, when such weights are used, the resulting
measures of model errors (i.e. standard error, standard error
of predictions, etc.) are no longer expressed in the original
scale of the data. To maintain the expressions of dispersion
in the original scale of the measurements, St-Pierre (2001)
suggested dividing each weight by the mean of all weights,
and to use the resulting values as weighing factors in the
analysis. Under this procedure, the average weight used
is algebraically equal to 1.0, thus resulting in expressions
of dispersion that are in the same scale as the original
data. The application of this technique is software dependent. In the Statistical Analysis System (SAS), for example,
weights are automatically rescaled so that their sums must
be equal to 1.
Weighing based on other criteria
Other weighing criteria have been suggested for the
weighing of observations, such as the power of an experiment to detect an effect of a size dened a priori, the
duration of an experiment, etc. The weighing scheme can
actually be based on an expert assessment, partially subjective, of the overall quality (precision) of the data. The
opinion of more than one expert may be useful in this
context. From a Bayesian statistical paradigm, the use of
subjective information for decision-making is perfectly
coherent and acceptable, as subjective probabilities are
often used to establish prior distributions in Bayesian
decision theory (De Groot, 1970). Traditional scientic objectiveness, however, may restrict the use of this weighing
scheme in scientic publications.
Predictably, the importance of weighing observations
decreases with the number of observations used in the
analysis, especially if the observations that would receive
a small weight have relatively small leverage values. For
example, we conducted a meta-analysis to quantify the effect
of concentrate intake on milk production and intake by dairy
cows (Sauvant et al., personal communication). The data,
consisting of 208 treatment means from 85 experiments that
had as a primary objective to study the effect of concentrate
intake, were analyzed with and without a weighing scheme
based on the reciprocals of the s.e. The resulting weights
ranged between 0.25 and 12.5. The functional structure of
the response was a quadratic function with xed study
effect. In the example, the dependent variable was DMI (kg/
day). The independent variable (predictor) was concentrate

DMI (CI, kg/day). For DMI, the estimated function using

unweighed observations was
DMI 16:7 0:41 0:64 0:09 CI
 0:018 0:004 CI2

RMSE 1:02:

The same function estimated using weighed observations

was:
DMI 17:5 0:41 0:48 0:09 CI
 0:012 0:005 CI2

RMSE 1:56:

Clearly, the regression coefcients of (2) and (3) are very

close to each other and do not lead to much difference in
the predicted response to CI.
Whether the analyst should weight the observations
based on the s.e. for each individual treatments or the
pooled s.e. from the studies is open to debate. There are
many reasons why the s.e. of each treatment within a study
can be different. First, the original observations themselves
could have been homoscedastic (homogeneous variance)
but the least-squares means would have different s.e. due
to unequal frequencies (e.g. missing data). In such a case,
it is clear that the weight should be based on the s.e. of
each treatment. Second, the treatments may have induced
heteroscedasticity, meaning that the original observations
did not have equal variances across sub-classes. In such
instances, the original authors should have conducted a test
to assess the usual homoscedasticity assumption in linear
models. The problem is that a lack of signicance (i.e.
P . 0.05) when testing the homogeneity assumption does
not prove homoscedasticity, but only that the null hypothesis (homogeneous variance) cannot be rejected at a
P , 0.05. In a meta-analytic setting, the analyst may deem
the means with larger apparent variance to be less credible
and reduce the weight of these observations in the analysis.
Unfortunately, most publications lack the information
necessary for this option.
Among the more subjective criteria available for weighing is the quality of the experimental design used in the
original study with regard to the meta-analytic objective.
Experimental designs have various trade-offs due to their
underlying assumptions. For example, the Latin square is
often used in instances where animal units are relatively
expensive, such as in metabolic studies. The double orthogonal blocking used to construct Latin square designs
can remove a lot of variation from the residual error. Thus
very few animals can be used compared to a completely
randomized design for an equal power of detecting treatment effect. The downside, however, is that the periods
are generally relatively short to reduce the likelihood of
a period by treatment interaction (animals in different
physiological status across time periods), thus reducing the
magnitude of the treatment effects on certain traits, such as
production and intake, for example. In those instances, the
analyst should legitimately weigh down observations from
1209

Sauvant, Schmidely, Daudin and St-Pierre

experiments whose designs limited the expression of the
treatment effects.

Statistical models
The independent variable can be either discrete or continuous. With binary data (healthy/sick, for example), generalized linear models (GLM) based on the logit or probit
link functions are generally recommended (Agresti, 2002).
Because of advances in computational power, the GLM has
been extended to include random effects in what is called
the generalized linear mixed model (GLMM). In its version
9, the SAS system includes a beta release of the GLIMMX
procedure to t these complicated models. In nutrition,
however, the large majority of the dependent variables
subjected to meta-analyses are continuous, and their analyses are treated at length in the remainder of this paper.
St-Pierre (2001) made a compelling argument to include
the study effect in all meta-analytic models. Because of the
severe imbalance in most databases used for meta-analyses, the exclusion of the study effect in the model leads to
biased parameter estimates of the effects of other factors
under investigation, and severe biases in variance estimates. In general, the study effect should be considered
random because it represents, in essence, the sum of the
effects of a great many factors, all with relatively small
effects on the dependent variable. Statistical theory indicates that these effects would be close to Gaussian
(normal), thus much better estimated if treated as random
effects. Practical recommendations regarding the selection
of the type of effect for the studies are presented in Table 2.
In short, the choice depends on the size of the conceptual
population, and the sample size (the number of studies in
the meta-analysis).
The ultimate (and correct) meta-analysis would be one
where all the primary (raw) data used to perform the
analyses in each of the selected publications were available
to the analyst. In such an instance, a large segmented
model that includes all the design effects of the original
studies (e.g. the columns and rows effects in Latin squares)
plus the effects to be investigated by the meta-analysis
could be tted by least-squares or maximum likelihood
methods. Although computationally complex, such huge
meta-analytic models should be no more difcult to solve
than the large models used by geneticists to estimate
the breeding values of animals using very large national

databases of production records. Raw data availability

should not be an issue in instances where meta-analyses
are conducted with the purpose of summarizing research at
a given research center. This, however, is very infrequent,
and meta-analyses are almost always conducted using
observations that are themselves summaries of prior experiments (i.e. treatment means). It seems evident that a metaanalysis conducted on summary statistics should lead to the
same results as a meta-analysis conducted on the raw data,
which itself would have to include a study effect because
the design effects are necessarily nested within studies (e.g.
cow 1 in the Latin square of study 1 is different from the
cow 1 of study 2), which itself would likely be considered
random. Thus, analytical consistency dictates the inclusion
of the study effect in the model, generally as a random
effect. The study effect will be considered random in the
remainder of this paper, with the understanding that
under certain conditions explained previously it should be
considered a xed effect factor.
Besides the study effect, meta-analytic models include
one or more predictor variables that are either discrete or
continuous. For clarity, we will initially treat the case of
each variable type separately, understanding that a model
can easily include a mixture of variables of both types, as
we describe later.

Model with discrete predictor variable(s)

A linear mixed model easily models this situation as follows:
Yijk m Si tj Stij eijk ;

where Yijk 5 the dependent variable, m 5 overall mean,

Si 5 the random effect of the i th study, assumed , iidN
(0, s2S), tj 5 the xed effect of the j th level of factor t,
Stij 5 the random interaction between the i th study
and the j th level of factor t, assumed , iidN (0, s2St),
and eijk 5 the residual errors, assumed , iidN (0, s2e).
eijk, Stij and Si are assumed to be independent random
variables.
For simplicity reasons, model (4) is written without weighing
the observations. The weights would appear as multiplicative factors of the diagonal elements of the error
variance-covariance matrix (Draper and Smith, 1998). Model
(4) corresponds to an incomplete, unbalanced randomized
block design with interactions in classic experimental

Table 2 Guidelines to establish whether an effect should be considered xed or random in a meta-analytic model*
Population
Case
Case
Case
Case

1
2
3
4

T is
T is
T is
T is

smalllarge
large
large

Experiment

Effect of t in the model

tT
t5T
tT
t 5 T, and t is very small

Fixed effect
Random effect
Should be xed but random works better
Should be random but xed may work better
(i.e. variance components can be poorly estimated)

*Adapted from Milliken (1999) for mixed models.

T represents the number of studies in the population (conceptual), and t is the number of studies in the meta-analysis.

1210

Meta-analyses in nutrition
research. The following SAS statements can be used to
solve this model:
PROC MIXED DATA Mydata CL COVTEST;
CLASSES study tau;
MODEL Y tau;

Yijk m Si tj Stij B1 Xij bi Xij Bj Xij eijk ;

8

RANDOM study study  tau;

LSMEANS tau;
5

RUN;

Standard tests of signicance on the effect of t are easily

conducted and least-squares means can be separated using
an appropriate mean separation procedure. Although it may
be tempting to remove the study effect from the model in
instances where it is not signicant (also called pooling of
effects), this practice can lead to biased probability estimations (i.e. nal tests on xed effects are conditional on tests
for random effects) and is not recommended. This is because
not being able to reject the null hypothesis of no study effect
(i.e. variance due to study is not signicantly different from
zero) is a very different proposition than proving that the
effect of the study is negligible. At the very least, the probability threshold for signicance of study should be much
larger than the traditional P 5 0.05 (say P , 0.25). Ideally, the
analyst should state before the analysis is performed what
size of estimated variance due to study should be considered
negligible, such as s2s o0:1 s2e.

Model with continuous predictor variable(s)

A linear mixed model also easily models this situation.
Yij B0 Si B1 Xij bi Xij eij ;

where Yij 5 the dependent variable, B0 5 overall (interstudy) intercept (a xed effect equivalent to m in (4)),
Si 5 the random effect of the i th study, assumed , iidN
(0, s2S ), B1 5 the overall regression coefcient of Y on X
(a xed effect), Xij 5 the value of the continuous predictor
variable, bi 5 the random effect of study on the regression
coefcient of Y on X, assumed , iidN (0, s2b ), and eij 5 the
residual errors, assumed , iidN (0, s2e ). Also, eij, bi and Si
are assumed to be independent random variables.
The following SAS statements can be used to solve this
model:
PROC MIXED DATA Mydata CL COVTEST;
CLASSES study;
MODEL Y X=SOLUTION;
RANDOM study study  X;
RUN;

Model with both discrete and continuous predictor

variable(s)
Statistically, this model is a simple combination of (4) and
(6) as follows:

Using a simple Monte Carlo simulation, St-Pierre (2001)

demonstrated the application of this model to a synthetic
dataset, showing the power of this approach, and the
interpretation of the estimated parameters.

where Bj 5 the effect of the jth level of the discrete factor t

on the regression coefcient (a xed effect).
The following SAS statements would be used to solve
this model:
PROC MIXED DATA Mydata CL COVTEST;
CLASSES study tau;
MODEL Y tau X tau  X;
RANDOM study study  tau study  X;
LSMEANS tau;
RUN;

In theory, (8) is solvable, but the large number of variance

components and interaction terms that must be estimated
in combination with the imbalance in the data makes it
often numerically intractable. In such instances, at least
one of the two random interactions must be removed from
the model.
In (4), (6) and (8), the analyst secretly wishes for the
interactions between study and the predictor variables to be
highly non-signicant. Recall that the study effect represents an aggregation of the effects of many uncontrollable
and unknown factors that differed between studies. A
signicant study by factor t interaction (Stij) in (4) implies
that the effect of t is dependent on the study, hence of
factors unaccounted for. Similarly, a signicant interaction
of study by X in (6) (bi Xij) indicates that the slope of the
linear relationship of Y on X is dependent on the study,
hence of unidentied factors. In such a situation, the
analysis produces a model that can explain very well the
observations, but predictions of future outcomes are generally not precise because the actual realization of a future
study effect is unknown. The maximum likelihood predictor
of a future observation is produced by setting the study and
the interaction of study with the xed effect factors to their
mean effect values of zero (McCulloch and Searle, 2001),
but the standard error of this prediction is very much
amplied by the uncertainty regarding the realized effect of
the future study.
When the study effect and its interaction with xed effect
is correctly viewed as an aggregation of many factors not
included in the model, but that differed across studies, the
desirability of including as many xed factors in the model
as can be uniquely identied from the data becomes
obvious. In essence, the xed effects should ultimately
make the study effect and its interactions with xed effects
predictors small and negligible. In such instances, the
resulting model should have wide forecasting applicability.
1211

Sauvant, Schmidely, Daudin and St-Pierre

Imagine, for example, that much of the study effect on body
weight gain of animal is in fact due to the large difference
in the initial body weight across studies. In such instances,
the inclusion of initial body weight as a covariate would
remove much of the study effect, and the diet effects
(as continuous or discrete variables) would be estimated
without biases, with a wide range of applicability (i.e. a
future prediction would require a measurement of initial
body weight as well as measurements of the other predictor
variables).
Whether one chooses (4) or (6) as a meta-model is
somewhat arbitrary when the predictor variable has an
inherent scale (i.e. is a measured number). The assumptions
regarding the relationship between Y and the predictor
variable are, however, very different between the two
models. In (4), the model does not assume any functional
form for the relationship. In (6), the model explicitly
assumes a linear relationship between the dependent and
the predictor variables. Different methods can be used to
determine whether the relationship should have a linear or
nonlinear structure.
>

>

>

The rst method consists in classifying observations into

ve sub-classes based on the quintiles for the predictor
variable, and performing the analysis according to (4) with
ve discrete levels of the predictor variable. Although
the selection of ve sub-classes is somewhat arbitrary,
there are substantive references in the statistical literature
indicating that this number of levels generally works well
(Cochran, 1968; Rubin, 1997). A visual inspection of the
ve least-squares means or the partitioning of the four
degrees of freedom associated with the ve levels of the
discrete variable into singular orthogonal polynomial
contrasts can rapidly identify an adequate functional form
to use for modeling the YX relationship.
The second method can be directly applied to the data, or
can be a second step that follows the identication of an
adequate degree for a polynomial function. Model (6) is
augmented with the square (and possibly higher order
terms) of the predictor variable. In the MIXED procedure
of SAS, this can be done simply by adding an X *X term
to the model statement. It is important to understand
that in the context of a linear (mixed or not) model, the
matrix representation of the model and the solution
procedure used are no different when X and X2 are in the
model compared to a situation where two different
continuous variables (say X and Z) are included in the
model. The problem, however, is that X and X2 are
implicitly not independent; after all, there is an algebraic
function relating the two. This dependence can result in a
large correlation between the two variables, thus leading
to possible problems of collinearity.
A third method can be used in more complex situations
where the degree of the polynomial exceeds two, or the
form of the relationship is sigmoid, for example. The
relationship can be modeled as successive linear segments,
an approach conceptually close to the rst method

1212

explained previously. Martin and Sauvant (2002) used this

method to study the variation in the shape of the lactation
curves of cows subjected to various concentrate supplementation strategies, using the model of Grosman
and Koops (1988) as its fundamental basis. Using this
approach, lactation curves were summarized by a vector of
nine parameter estimates, whose estimates could be
compared across supplementation strategies.
In (8), the interest may be in the effect of the discrete
variable t after adjusting for the effect of a continuous
variable X as in a traditional covariate analysis, or the
interest may be inverse, i.e. the interest is in the effect of
the continuous variable after adjusting for the effect of the
discrete variable. The meta-analyses of Firkins et al. (2001)
provide examples of both situations. In one instance, the
effect of grain processing (a discrete variable) on milk fat
content was being investigated while correcting for the
effect of DMI (a continuous variable). In this instance, the
interest was in determining the effect of the discrete variable. In another instance, the effects of various dietary
factors such as dietary NDF, DMI and the proportion of
forage in the diet (all continuous variables) on starch and
NDF digestibility, and microbial N synthesis were investigated, while correcting for the effect of the grain processing. In this case, the interest was much more towards the
effects of the continuous variables exempt from possible
biases due to different grain processing across experiments.

Accounting for interfering factors

Differences in experimental conditions between studies can
affect the treatment response. The nature of these conditions can be represented by quantitative or qualitative
variables. In the rst instance, the variable and possibly its
interaction with other factors can be added to the model
if there are sufcient degrees of freedom. The magnitude
of treatment response is sometimes dependent on the
observed value in the control group. For example, the milk
fat response in lactating cows to dietary buffer supplementation as a function of the milk fat of the control group
(Meschy et al., 2004). The response was small or nonexistent when the milk fat of control cows was near 40 g/l,
but increased markedly when the control cows had low milk
fat, possibly reecting a higher likelihood of sub-clinical
rumen acidosis in these instances.
When the intrastudy response depends on the levels of the
predictor variable, it is often because of the existence of a
nonlinear interstudy relationship. Applying model (6) with the
addition of a square term for the Xij to the data shown in
Figure 3 results in a relatively good quantication of the
relationship between chewing time and dietary NDF, as
shown in Figure 6. In this type of plot, it is important to adjust
the observations for the study effect, or the regression may
appear to poorly t the data because of the many hidden
dimensions represented by the studies (St-Pierre, 2001).
In instances where the interacting factor is discrete, the
examination of the sub-classes least-squares means can

Meta-analyses in nutrition
1000
10

800
Y = 55.9 + 29.4 X - 0.242 X2

700

(n = 216, nexp = 88, RMSE = 43.4)

Frequency

Chewing (min/d)

900

?
5

600
500

0
20

30

40
50
NDF (% DM)

60

70

Figure 6 Effect of dietary neutral detergent ber (NDF) content on

chewing activity in cattle. Data are from published experiments where the
NDF content of the diet was the experimental treatment. Observations
were adjusted for the study effect before being plotted, as suggested by
St-Pierre (2001). n 5 number of treatments; nexp 5 number of experiments.

clarify the nature of the interaction. For example, the effect

of a dietary treatment may be dependent on the physiological status of the animals used in the study. This physiological status can be coded using multiple dummy variables,
as explained previously.
Post-optimization analyses
As when tting conventional statistical models, numerous
analyses should follow the tting of a meta-analytic model.
These analyses are used to assess the assumptions underlying
the model, and to determine whether additional meta-analytic
models should be investigated (the heuristic process illustrated
in Figure 1).

Structure of residual variation

In (4), (6) and (8), the residuals (errors) are assumed
independent, and identically distributed from a normal
population with a mean of zero and a variance s2e. The
normality assumption can be tested using a standard x2
test, or a Shapiro-Wilks test, both available in the UNIVARIATE procedure of the SAS system. The residuals can
also be expressed as Studentized residuals, with absolute
values exceeding 3 being suspected as outliers (Tomassone
et al., 1983). In meta-analysis, the removal of a suspected
outlier observation should be done only with extreme
caution. This is because the observations in a meta-analysis
are the calculated outcomes (least-squares means) of
models and experiments that should themselves be nearly
free of the inuence of outliers. Thus, meta-analytic outliers
can be much more likely indicative of a faulty model than of
a defective observation. In addition, the removal of one
treatment mean as an observation in a meta-analysis might
be removing all the variation in the predictor variable for
the experiment in question, thus making the value of the
experiment in a meta-analytic setting nearly worthless. In
addition, the analyst should examine for possible intra- and

Standard errors of studies (min/d)

Figure 7 Frequency distribution of the residual standard errors of the

studies used in the meta-analysis of chewing activity in cattle.

interstudy relationships between the residuals and the

predictor variables.

Structure of study variation

When a model of the type described in (4) is being tted, it
is possible to examine each study on the basis of its own
residuals. For example, Figure 7 shows the distribution of
the residual standard errors for the different studies used in
the meta-analysis of chewing time in cattle (Figures 3
and 6). Predictably, the distribution is asymmetrical and
follows the law of Raleigh for the standard errors, while
variances have a x2 distribution. Studies with unusual
standard errors, say those with a x2 probability exceeding
0.999, could be candidates for exclusion from the analysis.
Alternatively, one could consider using the inverse of the
estimated standard errors as weights to be attached to the
observation before re-iterating the meta-analysis.
Other calculations such as leverage values, Cooks distances and other statistics can be used to determine the
inuence of each observation on the parameter estimates
(Tomassone et al., 1983).
Conclusion
Meta-analyses produce empirical models. They are invaluable for the synthesis of data that at rst may appear
scattered without much pattern. The meta-analytic process
is heuristic and implicitly allows returning to prior steps.
Extensive graphical analyses must be performed prior to the
parameterization of a statistical model to gain a visual
understanding of the data structure as well as to validate
data entries.
The increased frequency of meta-analyses published in
the scientic literature coupled with scarce funding for
research should create an additional need for scientic
journals to ensure that published articles provide sufcient
information to be used in a subsequent meta-analysis.
There may be a time when original data from published
articles will be available via the web in a standardized
format, a current practice in DNA sequencing research.
1213

Sauvant, Schmidely, Daudin and St-Pierre

Lastly, new meta-analytic methods should assist the
expansion of mechanistic modeling efforts of complex
biological systems by providing conceptual models as well
as a structured process for their external evaluation.
Acknowledgment
The authors are thankful to Philippe Faverdin for his comments
on an earlier draft of this paper.

References
Agresti A 2002. Categorical data analysis, 2nd edition. John Wiley & Sons,
Hoboken, New Jersey.
Bravo D, Sauvant D, Bogaert C and Meschy F III 2003. Quantitative aspects of
phosphorus excretion in ruminants. Reproduction, Nutrition, Development 43,
285300.
Cochran WG 1968. The effectiveness of adjustment by subclassication in
removing bias in observational studies. Biometrics 24, 295313.
De Groot MH 1970. Optimal statistical decision. Mc Graw-Hill, New York.
Draper NR and Smith H 1998. Applied regression analysis. John Wiley & Sons,
Inc., New York.
Euge`ne M, Archime`de H and Sauvant D 2004. Quantitative meta-analysis on
the effects of defaunation of the rumen on growth, intake and digestion in
ruminants. Livestock Production Science 85, 8191.
Firkins JL, Eastridge ML, St-Pierre N and Noftsger SM 2001. Invited: Effects of
grain variability and processing on starch utilization by lactating dairy cattle.
Journal of Animal Science 79(E. Suppl.), E218E238.
Glass GV 1976. Primary, secondary and meta-analysis of research. Education
Research 5, 38.
Grosman M and Koops WJ 1988. Multiphasic analysis of lactation curves in
dairy cattle. Journal of Dairy Science 71, 15981608.
Lovatto PA and Sauvant D 2002. Meta-analyse et modelisation de lingestion
volontaire chez le porc. Journees de la Recherche Porcine, Paris (FRA) 2002/02/
0507, 129134.
Mantel N and Haenszel W 1959. Statistical aspects of the analysis of data from
retrospective studies of disease. Journal National of Cancer Institute 22,
719748.

1214

Martin O and Sauvant D 2002. Meta-analysis of input/output kinetics in

lactating dairy cows. Journal of Dairy Science 85, 33633381.
McCulloch CE and Searle SR 2001. Generalized, linear, and mixed models. John
Wiley & Sons, Inc., New York.
McPhee MJ, Oltjen JW, Famula TR and Sainz R 2006. Meta-analysis of factors
affecting carcass characteristics of feedlot steers. Journal of Animal Science 84,
31433154.
Meschy F, Bravo D and Sauvant D 2004. Analyse quantitative des reponses des
vaches laitie`res a` lapport de substances tampon. INRA Productions Animales
17, 1118.
Milliken GA 1999. Mixed models. Seminar to the Cleveland Chapter of the
American Statistical Association, May 10.
Offner A and Sauvant D 2004. Prediction of in vivo starch digestion in cattle
from in situ data. Animal Feed Science and Technology 111, 4156.
Offner A, Bach A and Sauvant D 2003. Quantitative review of in situ starch
degradation in the rumen. Animal Feed Science and Technology 106, 8193.
Rico-Gomez M and Faverdin P 2001. La nutrition proteique modie lingestion
des vaches laitie`res: analyse bibliographique. Rencontres Recherches
Ruminants 8, 285288.
Rubin DB 1997. Estimating causal effects from large data sets using propensity
scores. Annals of Internal Medicine 127, 757763.
Sauvant D 1992. La modelisation systemique en nutrition. Reproduction,
Nutrition, Development 32, 217230.
Sauvant D 1994. Modelling homeostatic and homeorhetic regulations in
lactating animals. Livestock Production Science 39, 105113.
Sauvant D and Martin O 2004. Empirical modelling through meta-analysis vs
mechanistic modelling. In Nutrient digestion and utilization in farm animals:
modelling approaches (ed. E Kebreab). CAB International, London.
St-Pierre NR 2001. Invited review: integrating quantitative ndings from
multiple studies using mixed model methodology. Journal of Dairy Science 84,
741755.
St-Pierre NR and Glamocic D 2000. Estimating unit costs of nutrients from
market prices of feedstuffs. Journal of Dairy Science 83, 14021411.
Schmidely P, Glasser F, Doreau M and Sauvant D 2008. Digestion of fatty acids
in ruminants: a meta-analysis of ows and variation factors. 1. Total fatty
acids. Animal 2, 677690.
Tomassone R, Lesquoy E and Millier C 1983. La regression: nouveau regard sur
une ancienne methode statistique. In Actualites scientiques et agronomiques
de lINRA (ed. Masson). INRA, Paris.