Sunteți pe pagina 1din 19

The information-energy model is a theory which purports to explain changes in

bureaucratic development for developing nations in terms of changes in ecological


variables. This research note attempts to test the validity of the model and to
assess its explanatory power through aggregate data analysis and cross-national
causal modeling techniques. The results indicate that though ecological variables
account for some variation in bureaucratic development, the overall explanatory
power of the model is not what was initially hypothesized. Of all the paths in the
causal model leading to bureaucratic development, the direct linkage between
information and bureaucratic development is the strongest.

TESTING THE
INFORMATION-ENERGY MODEL

WILLIAM M. BERENSON
Boston University Law School

The relation between bureaucratic development and ecologi-


cal variables is a topic which has generated a considerable
amount of inquiry and theoretical speculation within the
field of comparative administration. Rarely, however, have
data been systematically introduced in order to operation-
alize and test the numerous schemas proposed by scholars
and armchair theorists examining this relationship (Sigelman,
1972). Hence, most schemas relating ecological variables to
bureaucratic development remain merely paper theories.

AUTHOR’S NOTE: Acknowledgments are due to John T. Dorsey, Lee Sigelman,


and Robert D. Bond who commented on earlier drafts of this research note.

[139]

Downloaded from aas.sagepub.com at PENNSYLVANIA STATE UNIV on September 18, 2016


[140]ADMINISTRATION AND SOCIETY / AUGUST 1977

They have yet to be either verified or nullified by hypothesis


testing and systematic data analysis. This research note is an
attempt to shed some light on the validity of one of these
schemas, John T. Dorsey’s Information-Energy Model (1962,
1963), by operationalizing the key variables and systemati-
cally testing causal relations suggested from the model.

THE INFORMATION-ENERGY MODEL

Dorsey’s model is the logical extension of the work of two


anthropologists, Leslie White (1949) and Fred Cottrell
(1955), who in the late 1940s and early 1950s attempted to
explain cultural and social development of societies as a
function of their energy and information resources. Changing
the dependent variable from sociocultural development to
bureaucratic development, Dorsey has combined various
aspects of the work of both White and Cottrell with a

generous sprinkling of Karl Deutsch’s cybernetics (Dorsey,


1962). The result of this synthesis is the information-energy
model.
The information-energy model focuses on the interaction
of three ecological variables-energy, information, and energy
conversion-and their relation to a dependent variable, bu-
reaucratic development. The network of relations between
these four variables is graphically depicted in Figure 1.
Energy is the independent variable and is hypothesized to
affect bureaucratic development through the intervening vari-
ables, information and energy conversion. The solid lines in
Figure 1, E―~-1―~-C―~B, represent the developmental
sequence of the causal relations between these variables.
The more explicit interactions of these variables within the
model may be briefly summarized as follows: within the
social and physical environment of each nation state, there
exists a certain quantity of potential energy. This energy may
be in the form of water for transportation and power, fossil

Downloaded from aas.sagepub.com at PENNSYLVANIA STATE UNIV on September 18, 2016


Berenson / TESTING INFORMATION-ENERGY MODEL [141]

FIGURE 1
The Information-Energy Model :
Hypothesized Causal Relations

fuels, or aworking population of considerable magnitude.


Similarly, information exists in the environment of the
nation state. This information is usually converted from the
existing sources of potential energy within the state; how-
ever, it may be imported in the form of technology from
other states as well. Dorsey posits that as the known sources
of energy and the level of information rise within a nation
state, they stimulate energy conversion or economic develop-
ment. The magnitude and scope of energy conversion is thus
positively related to the quantity of energy and information
circulated within the polity (Dorsey, 1962, 1963).
Of the two variables, energy and information, the latter is
markedly more critical in stimulating increased rates of
energy conversion than the former. Dorsey theorizes that a
social system’s overall level of development is &dquo;an expression
of its energy conversion levels ... [which] depend upon the

Downloaded from aas.sagepub.com at PENNSYLVANIA STATE UNIV on September 18, 2016


[142] ADMINISTRATION AND SOCIETY / AUGUST 1977

information available to, stored in, and circulating within the


system&dquo; (Dorsey, 1963: 321). The quantity and type of
information available determine the quantity and methods of
energy release, expansion, and consumption within a society.
The final linkage between energy conversion and bureau-
cratic development is perhaps best summarized by Dorsey
himself:

As the levels of energy conversion rise, the political, social and


economic structures undergo transformations. The structural pat-
terns and processes sufficient for &dquo;low energy&dquo; [underdeveloped]
societies must be replaced by or transformed into different pat-
terns of structures and processes in societies of rising energy
levels. [Dorsey, 1963: 320]

One of these structural patterns or processes to be trans-


formed is the bureaucracy. As the levels of economic devel-
opment or energy conversion increase, the bureaucracy, like
society, becomes more differentiated and modernized.
From the above description of the information-energy
model, we have extracted two central propositions which,
when tested as hypotheses, should shed some light on the
validity of the entire schema. These are: Pi-the three vari-
ables, energy, information, and energy conversion, together
explain most of the variance in bureaucratic development
among developing nations ;2 and P2 -the principal configura-
tion of paths through which energy influences bureaucratic
development is the developmental sequence outlined in Fig-
ure 1. The remainder of this research note is devoted to

testing the validity of these two propositions.

METHODOLOGY

1. CAVEATS

In attempting to operationalize and collect data pertaining


to the four variables in the information-energy model, several

Downloaded from aas.sagepub.com at PENNSYLVANIA STATE UNIV on September 18, 2016


Berenson / TESTING INFORMATION-ENERGY MODEL[143]

serious problems were encountered. Some of these difficul-


ties were resolved more satisfactorily than others.
The operationalization of the ecological variables posed a
relatively minor problem. The two articles in which Dorsey
lays out the information-energy model served as a helpful
guide for constructing operational indices of these ecological
variables. Unfortunately, at the time of this writing, there are
no adequate indicators of bureaucratic development and the

possibility of constructing a satisfactory index for this vari-


able, particularly for measuring bureaucratic development in
the developing nations, remains remote.
The difficulty in obtaining data pertaining to the four
variables for a universe of nations large enough to allow for
meaningful hypothesis testing was the greatest problem
encountered in this project. To convincingly test the type of
causal relationships set forth in the information-energy
model, longitudinal data or lagged data are preferred (Sigel-
man, 1971; Burrowes, 1970; Pride, 1971: 363). However,
longitudinal data for most of the information-energy vari-
ables within an acceptable universe of nations are not avail-
able. Consequently, in collecting our data, we had to resort
to cross-sectional indicators which are admittedly inferior for
inferring the validity of the hypothesized causal sequences in
the model. Another problem with the type of data employed
in this analysis is its questionable reliability. By now, it is
common knowledge that data emanating from most national

governments, particularly the developing nations, are often


inflated for political reasons and are rarely an honest indica-
tor of existing conditions (Schench, 1966).
Certainly the problems are not unique to this project. With
few exceptions, they pervade most cross-national research
conducted in our field. Nevertheless, since these problems
have not been resolved in this article, the reader is urged to
bear the above caveats in mind when drawing conclusions
from this and other cross-national studies of bureaucratic
development.

Downloaded from aas.sagepub.com at PENNSYLVANIA STATE UNIV on September 18, 2016


[144] ADMINISTRATION AND SOCIETY / AUGUST 1977

2. DATA

Data for 28 variables posited to be operational indicators


of the four variables in the information-energy model were
collected for a universe of 56 nations.3 Twenty-six of the
variables were subsequently factor analyzed and indices of
energy, energy conversion, and information were constructed
from the best factor emerging from the results of the factor
analysis.4 A factor score on each of these indices was com-
puted for each of the 56 nations in our universe and these
factor scores were used as the data in the regression equations
employed for testing the causal relations in the models (see
Appendix).
3. OPERATIONALIZATION

Dorsey (1963: 320) defines energy in a given social system


as not only &dquo;fuels, water power, etc. used to run machines
but also the energy expended in the physical and other
activities of the people as well as domesticated animals used
for human purposes.&dquo; Using Dorsey’s conceptual definition
as a guide, we collected data for the following indicators:

density, population growth, working-age population in the


working force, food supply, and energy potential. These
indicators were then factor analyzed in order to construct an
index of energy within a society. Here we assumed that a
factor which exhibited high positive loadings for the indica-
tors energy potential, food supply, and percent working age
population in the work force, and negative loadings on den-
sity and population growth6 would provide us with an ade-
quate weighted index of energy. The results of our factor
analysis appear in Table 1. Factor B was selected as the
energy index since it came closest to meeting our
assumptions.
Information is described by Dorsey (1963: 320) as a
kinetic process which requires &dquo;the systematic abstraction,

Downloaded from aas.sagepub.com at PENNSYLVANIA STATE UNIV on September 18, 2016


Berenson / TESTING INFORMATION-ENERGY MODEL [145]

TABLE 1
Factor Analysis of Energy Indicators

organization, accumulation, and transmission of increasing


amounts of information ... organized knowledge ... about
empirical reality.&dquo; Keeping Dorsey’s language definition in
mind, we collected data for the following variables: percent
of physicians and dentists, percent of population urban,
literacy rates, newspaper circulation, radios per capita, cin-
ema attendance per capita, domestic mail flows, and foreign
mail flows. These variables were factor analyzed and it was
assumed that a factor on which all of these variables loaded
strongly and positively would be a satisfactory indicator of
information. The results of our factor analysis presented in
Table 2 indicate that factor A came closest to meeting our
assumptions. Consequently, it was selected as our weighted
index of information in developing societies.
Like information, energy conversion is defined by Dorsey
as a kinetic process; it involves the releasing, expanding, and

consuming of energy. In his writings, Dorsey (1962: 40;


1963: 321 ) suggests that scientific agriculture, electric power,
steam engines, and gasoline motors may be used as indicators
of energy conversion within a society. In operationalizing
energy conversion, we collected data for indicators related to
those suggested by Dorsey. These were: railroad density,

Downloaded from aas.sagepub.com at PENNSYLVANIA STATE UNIV on September 18, 2016


[146] ADMINISTRATION AND SOCIETY / AUGUST 1977

TABLE 2
Factor Analysis of Information Indicators

railroad kilometers per capita, energy consumption per cap-


ita, commercial energy consumption, electric energy, hydro-
electric generation, steel consumption, fertilizer consump-
tion, percent employed in agriculture, and GNP per capita.
These indicators were factor analyzed in order to construct a
weighted index on energy conversion. We assumed that the
factor on which all indicators except percent employed in
agriculture exhibited positive and strong loadings would be
the factor most suitable for forming a valid energy conversion
index. We also assumed that percent employed in agriculture
would load highly negative on this factor since a large pri-
mary sector is usually associated with low energy conversion
societies. Given these assumptions and the results of our
factor analysis, we chose factor A in Table 3 as the most
suitable weighted index for energy conversion.

Downloaded from aas.sagepub.com at PENNSYLVANIA STATE UNIV on September 18, 2016


Berenson / TESTING INFORMATION-ENERGY MODEL[147]

TABLE 3
Factor Analysis of Energy Conversion Indicators

In attempting to operationalize bureaucratic development,


we were left sitting on the horns of several uncomfortable
dilemmas. Dorsey’s description of the information-energy
model provided few clues for finding a suitable index for this
variable. Our first dilemma was to choose between designing
a new index of bureaucratic development or using the

existing scales within the discipline. Initially the former strat-


egy seemed preferable to the latter. The two existing scales
for measuring bureaucratic development are beset with a
number of problems. The first, the Banks and Textor (1963)
index, has been criticized on counts of multidimensionality
and unreliability (Blank, 1965: 6). The second, the Adelman
and Morris (1967) administrative effectiveness index, is per-
haps more discrete in detecting differences among countries
than the Banks and Textor measure but its reliability is still
suspect.7 Attempting to avoid using these scales because of

Downloaded from aas.sagepub.com at PENNSYLVANIA STATE UNIV on September 18, 2016


[148] ADMINISTRATION AND SOCIETY / AUGUST 1977

their methodological deficiencies, we set out to construct a ’


new index of bureaucratic development; however, our efforts
in this enterprise were soon thoroughly thwarted by data
constraints.8
Our first dilemma having been resolved by default, we
were next forced to choose between the Banks and Textor
and the Adelman and Morris indices for our index of bureau-
cratic development. Though the Adelman and Morris scale
may be more accurate than the Banks and Textor, index
scores are available for a larger universe of nations using the
former index rather than the latter. Since the probability of
discovering statistically significant relations between bureau-
cratic development and our three ecological variables
increases with the N of our sample, we ultimately decided to
use the Banks and Textor indicator rather than Adelman and
Morris’ as our independent variable.9

4. TESTING THE PROPOSITIONS

Multiple regression analysis and causal modeling tech-


niques were used to test the validity of the two central
propositions extracted from the information-energy model.
Our results indicate that neither of these propositions can be
assumed true for the developing countries in our 56 nation
universe. 10
In order to ascertain the validity of our first proposition-
that the three ecological variables, energy, information, and
energy conversion, together explain most of the variance in
bureaucratic development among developing nations-we
regressed the Banks and Textor Bureaucratic Development
Index on our three independent variables. The regression
analysis yielded a multiple R of .5859 so that together, the
three variables were found to explain a mere 34.38% of the
total variance in bureaucratic development. 11 With 65.62% of
the variance left unexplained, we concluded that the three
ecological variables in the information energy model fail to

Downloaded from aas.sagepub.com at PENNSYLVANIA STATE UNIV on September 18, 2016


Berenson / TESTING INFORMATION-ENERGY MODEL[149]

FIGURE 2
Correlation Coefficients and Paths Between
the Four Variables in the Information-Energy Model

offer an adequate explanation of bureaucratic development


in the &dquo;third world.&dquo;
Our second proposition-that the principal configuration
of paths through which energy influences bureaucratic devel-
opment is the developmental sequence E―~-1―~C―~-B―
was tested by standard causal modeling procedures. The
Blalock method of using product moment correlations to fit
the data to prediction equations was first employed to elimi-
nate spurious relations between the variables (Alker, 1965;
Blalock, 1971: 18-32). Next, standardized multiple regression
coefficients, multiple Bs (beta weights) were calculated and
substituted as path coefficients in the model. By summing
the beta weights of the component paths within each possible
root path, it was possible to determine which of the possible
root paths between the four variables is the strongest for
influencing bureaucratic development (Smith, 1969; McCrone
and Cnudde, 1967).
Figure 2 depicts the correlations between the four vari-
ables in the information-energy model along with a schematic
representation of all possible paths. If the second proposition
is true, the paths represented by the dotted lines are either
weak or spurious and those paths connected by the solid
lines-the developmental sequence-are not spurious.
Using prediction equations according to the methods pre-
scribed by Blalock (Alker, 1965: 122; Blalock, 1971), we

Downloaded from aas.sagepub.com at PENNSYLVANIA STATE UNIV on September 18, 2016


[150] ADMINISTRATION AND SOCIETY / AUGUST 1977

TABLE 4
Prediction Equations for
the Information-Energy Model

were able to cast doubt upon the validity of our second


proposition. The results in Table 4 show that of the three
component paths represented by dotted lines, only the direct
path E―~B is unequivocally spurious. Among the paths
represented by the solid line, there is a problem with path
E 0 1. Though this path is not statistically spurious accord-
ing to the above table, it makes little substantive sense to
keep it within the model. In our earlier discussion of the
’It

Downloaded from aas.sagepub.com at PENNSYLVANIA STATE UNIV on September 18, 2016


Berenson / TESTING INFORMATION-ENERGY MODEL[151]

FIGURE 3
The Information-Energy Model
Revised: Paths and Beta Weights

information-energy model, we indicated that the model


assumed energy to be positively related to information; how-
ever, when the relationship is controlled for the variable
energy conversion, the partial correlation between informa-
tion and energy drops from a low positive to a low negative
figure. This becomes readily apparent when one remembers
that the difference figure in Table 4 for equation 4 is actually
the numerator of the partial coefficient ri -ec- Therefore,
we decided that the path rei, while not statistically insignifi-

cant, is substantively incongruous with the model as origi-


nally described in this paper. With the direct paths E―~B
and E 0removed as either statistically or substantively
spurious, the information-energy model takes on the form of
the causal model depicted in Figure 3.
Calculating path coefficients according to the rules set
down by Sewall Wright in the 1920s (Wright, 1934; Smith,
1969; Turner and Stevens, 1971), we were able to deduce
which of the paths in the revised version of the information-
energy model is the strongest-i.e., the strongest path effects
the greatest change in bureaucratic development. From our
analysis of the path coefficients (beta weights) depicted in
Figure 3, we concluded that the strongest path in the revised
model is the direct linkage between information and bureau-
cratic development. 12 For every unit increase in information,
bureaucratic development increases .332 units. This is more
than a .140 increase over the amount of change effected by

Downloaded from aas.sagepub.com at PENNSYLVANIA STATE UNIV on September 18, 2016


[152] ADMINISTRATION AND SOCIETY / AUGUST 1977

TABLE 5
Path Analysis of the
Revised Information-Energy Model

either of the other two possible paths, E―~-C―~-B and


I―~&dquo;C―~-B. The full results of our path analysis are pre-
sented in Table 5.

CONCLUSIONS

As operationalized in this research note, the key indepen-


dent variables in the information-energy model are not as
predictive of bureaucratic development as initially posited.
Together they leave a majority of the variance in bureaucratic
development unexplained-a finding which refutes any pre-
sumptions as to the model’s worth as a total explanatory
schema. Nonetheless, the results indicate that though ecologi-
cal variables do not exercise the quantum of explanatory
influence on bureaucratic development proposed by the
model, they do explain enough variance to mitigate against
discounting their impact altogether.
The strongest path in the information-energy model is not
the developmental sequence posited at the outset of this
research note. Instead, the strongest path is the direct linkage
between information and bureaucratic development. This

Downloaded from aas.sagepub.com at PENNSYLVANIA STATE UNIV on September 18, 2016


Berenson / TESTING INFORMATION-ENERGY MODEL[153]

fact, combined with our findings that the direct linkages


E―~-B and E 0are spurious, allows us to infer that the
energy level in a society is not critical to bureaucratic devel-
opment. In fact, our data indicate that a society with little
natural energy potential can both develop a modem bureau-
natural energy potential can both develop a modern bureau-
cracy and sustain energy conversion activities as long as
converted from domestic sources of energy, but can be
imported from other societies in the form of foreign tech-
nical and financial assistance. This finding is generally sup-
portive of the greater theoretical importance given to
information over energy by Dorsey (Dorsey, 1963: 320).
Finally, the relatively low explanatory power of the infor-
mation-energy model as represented in this note may be
attributed to several factors. One is the global definitional
breadth of the ecological variables. Though suitably concep-
tualized for providing a sound framework for a more qualitat-
ive case by case analysis from which theoretical generaliza-
tions can be drawn, they are perhaps too inclusive or
connotative to be of much use in the more expedient type of
theory testing attempted here. Another explanation for the
rather disappointing results may lie in the data and type of
modeling apparatus employed. The recursive nature of the
causal model may not do justice to the more synergistic
relationship between bureaucratic development and the
ecological variables suggested by Dorsey. Moreover, the afore-
mentioned weaknesses in the data and indices of bureaucratic
development utilized in this analysis leave open the possi-
bility that more definitive and encouraging results may be
derived from subsequent replications of this test as more
reliable data and indices become available.

Downloaded from aas.sagepub.com at PENNSYLVANIA STATE UNIV on September 18, 2016


[154] ADMINISTRATION AND SOCIETY / AUGUST 1977

NOTES

1. An ecological theory of bureaucratic development assumes that change in


the bureaucracy is somehow related to change in other subsystems in the society
such as the economy, social structures, and aspects of the political system.
Ecological theory is a systematic effort to relate public administration to the
environment (Heady, 1966: 24). Examples of ecological theories are Riggs
(1959), Eisenstadt (1959), and Weber (Gerth and Mills, 1946). For a more
detailed discussion of ecological models, see Sigelman (1971: 19-20, 4-8).
2. Since Dorsey (1962) indicates that his model is most applicable for
examining bureaucratic development in the developing nations, we have limited
our analysis only to nations classified as developing. Here we define developing
nations operationally as those with GNP’s per capita below $550.
3. The data used in this analysis were gathered from Ginsberg (1961),
Russett et al. (1964), and Banks and Textor (1963). Most of the data for the 28
variables were available in these sources. When data on important variables were
not available for a country, the country with the missing data was either
eliminated from our universe or a datum for the missing indicator was estimated
from that country’s performance on other related indicators and general informa-
tion about that country. This data estimation procedure, however, was kept to a
minimum. Data were finally accumulated for a universe of 56 developing nations.
4. The principal components method for extracting factors was used here
instead of the centroid and cluster methods because it offers a unique least squares
solution. The first factor extracted under the principal components analysis
accounts for the greatest variance in all the indicators; the next factor accounts
for the greatest proportion of residual variances and so forth. The varimax
method was then employed to rotate the extracted factors orthogonally in order
to insure their mutual independence and to improve the loadings (Palumbo, 1969:
285-305).
5. Factor scores were calculated with a program which uses least squares
methods to fit the scores to the data (Rummel, 1970: 437-441).
6. Population growth rate and density appear to be negative energy indi-
cators for the underdeveloped nations in our sample. The data suggest that highly
dense and rapidly growing population concentrations are a drain on other energy
resources. Both population growth rate and density are negatively correlated with
each other at -.30. Population growth rate is negatively correlated with all the
energy indicators in the factors. Density exhibits a .42 correlation with working
age population, but is insignificantly associated with food supply and energy
potential at .07 and .08 respectively.
7. Scores on the Banks and Textor index (1963: 112) range from 1-4 and
include the following scale types:

(4) Modern (generally effective and responsible civil service or the equivalent
performing in a functionally specific and nonascriptive context);
(3) Semimodern (largely "rationalized" bureaucratic structure of limited
efficiency because of shortages of skilled personnel, inadequacy of re-

Downloaded from aas.sagepub.com at PENNSYLVANIA STATE UNIV on September 18, 2016


Berenson / TESTING INFORMATION-ENERGY MODEL [155]

cruitment or performance criteria, excessive intrusion by nonadministra-


tive organs or partially noncongruent social institutions);
(2) Post Colonial Transitional (largely rationalized excolonial bureaucratic
structure in process of personnel nationalization and adaptation to the
services of restructuring of autochthonous social institutions); and
(1) Traditional (largely nonrationalized bureaucratic structures performing in
the context of an ascriptive and deferential stratification system).
The Adelman and Morris index (1967: 77-78) ranks nations on a continuum
of bureaucratic effectiveness from 01 to 100. The index actually uses letter
rankings A+ to C- but each letter can be assigned a numerical value. The nine
different scale types and their numerical scores are as follows:

A. Countries in which public administration is efficient with a well trained


civil service, no instability of existing policy, no widespread corruption
and no marked bureaucratic inefficiency.
High A+ 100 Medium A 90 Low A- = 80
= = = = =

B. Countries in which public administration is marked by considerable


inefficiency. These countries have a permanent body of administrators
and common corruption but the corruption does not interfere with
government functioning.
High =
B+ =
60 Medium =
B =
50 Low =
B- =
40
C. Countries characterized by extreme bureaucratic inefficiency. Here cor-
ruption is widespread and there is serious instability of policy.
High = C+ 20 Medium C 10 Low C- 01
= = = = =

8. This index of bureaucracy was to be a composite of several indicators


which we hypothesized to be highly characteristic of a modem public administra-
tive system. These indicators were: government expenditures as a percent of GNP,
government revenue as a percent of GNP, Social Security as a percent of GNP,
and the Banks and Textor scores. Though we could find data for these indicators
in the developed nations, they were not available for a sufficiently large universe
of developing nations.
9. The Banks and Textor index is an ordinal scale but in using it as the
dependent variable in a regression, we are assuming it equivalent to an interval
scale. While in the past this has been regarded as methodologically fallacious, it
has recently been shown by statisticians that an ordinal scale can be used as
interval data without distortion of results (Labovitz, 1970).
10. Here a note of caution is appropriate. Scattergrams of the dependent
variable with the three independent variables reveal that the assumption of
homoscedasticity necessary for regression analysis is far from perfectly met by
our data.
11. The regression was also calculated for a universe of 43 nations using the
Adelman and Morris index. Here the three independent variables succeeded in
explaining only 13.09% of the variance at the .07 level of significance. In this
equation, homoscedasticity was not a potential problem.

Downloaded from aas.sagepub.com at PENNSYLVANIA STATE UNIV on September 18, 2016


[156]ADMINISTRATION AND SOCIETY / AUGUST 1977

12. The direct links E → I and E → B similarly disappear when the Adel-
man and Morris index is used as the dependent variable for a universe of 43
nations. There are, however, some striking differences between the 56 nations and
the 43 nation model. In the latter, the strongest linkage is I → C → B which
accounts for .1973 of the .3888 total units change in B. Whereas the direct link
I → B is the strongest for the former, it is spurious for the latter. Nevertheless,
the direct effects of energy upon both bureaucratic development and information
levels are revealed to be nonexistent in each of these two models.

REFERENCES

ADELMAN, I. and C. T. MORRIS (1967) Society, Politics, and Economic


Development: A Quantitative Approach. Baltimore: Johns Hopkins Press.
ALKER, H. (1965) Mathematics and Politics. London: Macmillan.
BANKS, A. S. and R. B. TEXTOR (1963) A Cross Polity Survey. Cambridge,
Mass: MIT Press.
BLALOCK, H. (1971) "Four-variable causal models and partial correlations," pp.
18-32 in H. Blalock (ed.) Causal Models in the Social Sciences. Chicago:
Aldine-Atherton.
BLANK, B. (1965) "A proposal for a statistical approach to comparative admin-

istration : the measurement of national bureaucracies." Indiana: CAG Occa-


sional Papers.
BURROWES, R. (1970) "Multiple time series analysis of national level data."
Comparative Pol. Studies 2 (January): 465-480.
COTTRELL, W. F. (1955) Energy and Society: The Relation Between Energy,
Social Change, and Economic Development. Westport, Ct.: Greenwood.
DORSEY, J. (1963) "The bureaucracy and political development in Viet Nam,"
pp. 318-359 in J. LaPalombara (ed.) Bureaucracy and Political Development.
Princeton: Princeton Univ. Press.
---

(1962) "An information-energy model," pp. 37-57 in F. Heady and S.


Stokes (eds.) Papers in Comparative Administration. Ann Arbor: Institute of
Public Administration, University of Michigan.
EISENSTADT, S. N. (1959) "Bureaucracy, bureaucratization, and debureau-
cratization." Admin. Sci. Q. 4 (December): 302-320.
GERTH, H. H. and C. W. MILLS (1946) From Max Weber: Essays in Sociology.
New York: Oxford Univ. Press.
GINSBERG, N. (1961) Atlas of Economic Development. Chicago: Univ. of
Chicago Press.
HEADY, F. (1966) Public Administration: A Comparative Perspective. Engle-
wood Cliffs, N.J.: Prentice-Hall.
LABOVITZ, S. (1970) "The assignment of numbers to rank order categories."
Amer. J. of Sociology 35 (June): 515-524.

Downloaded from aas.sagepub.com at PENNSYLVANIA STATE UNIV on September 18, 2016


Berenson / TESTING INFORMATION-ENERGY MODEL [157]

McCRONE, D. J. and C. F. CNUDDE (1967) "Toward a communications theory


of democratic political development: a causal model." Amer. Pol. Sci. Rev. 61
(March): 72-79.
PALUMBO, D. J. (1969) Statistics in Political and Behavioral Science. New York:
Appleton-Century-Crofts.
PRIDE, R. (1971) "An alternative approach to quantitative historical data."
Comparative Pol. Studies 4 (October): 373-368.
RIGGS, F. (1959) "Agraria and industria: toward a typology of comparative
administration," pp. 23-116 in W. Siffin (ed.) Toward the Comparative Study
of Public Administration. Bloomington: Indiana Univ. Press.
RUMMEL, R. J. (1970) Applied Factor Analysis. Evanston: Northwestern Univ.
Press.
RUSSETT, B. et al. (1964) Yale Handbook of World Indicators. New Haven: Yale
Univ. Press.
SCHENCH, E. K. (1966) "Cross-national comparisons using aggregate data: some
substantive and methodological problems," in R. L. Merritt and S. Rokkan
(eds.) Comparing Nations: The Use of Quantitative Data in Cross National
Research. New Haven: Yale Univ. Press.
SIGELMAN, L. (1972) "Administrative development and dominance: a new test
of the imbalance thesis." Nashville: Department of Political Science, Vander-
bilt University. (unpublished)
(1971) "Modernization and the political system: a critique and preliminary
———

analysis." Sage Professional Papers in Comparative Politics 01-016. Beverly


Hills, Calif.: Sage.
SMITH, A. K. (1969) "Socioeconomic development and political democracy: a
causal analysis." Midwest J. of Pol. Sci. 30 (February): 95-125.
TURNER, M. E. and C. STEVENS (1971) "The regression analysis of causal
paths," pp. 75-100 in H. Blalock (ed.) Causal Models in the Social Sciences.
Chicago: Aldine-Atherton.
WHITE, L. A. (1949) The Science of Culture. New York: Farrar, Straus.
WRIGHT, S. (1934) "The method of path coefficients." Annals of Mathematical
Statistics 5: 161-215.

William M. Berenson is presently at Boston University Law School. Dr.


Berenson did public administration related research in Uruguay under a
FAFP grant in 1972-1973. He has taught political science at Fisk Univer-
sity, was a research fellow at M.I. T.’s Center for International Studies, and
has worked in Washington, D.C. as a consultant in public administration
related areas. Dr. Berenson has published in several social science journals.

Downloaded from aas.sagepub.com at PENNSYLVANIA STATE UNIV on September 18, 2016

S-ar putea să vă placă și