UNIT OF ANALYSIS
The things that researchers are
comparing or examining are referred to as the units of analysis, or the units to be
analyzed. Social research looks for patterns when comparing things one to
The most frequent unit of
analysis is the individual, suggesting that researchers look for patterns among a
collection (perhaps a sample) of individuals.
Research can also be conducted
in which a pattern is sought among a collection of groups; the group would be the unit of
For example, like Durkheim, one
might try to determine what social factors are linked to the variation in suicide rates among nations or regions.
One can also look for patterns
among things like newspaper stories, advertisements, a category of social interaction,
social events, or speech utterances. In this case the unit of analysis would be what Earl
Babbie has called social artifacts.
In Search of `the
Appropriate' Unit of Analysis for Sociocultural Research
Eugene Matusov, University of Delaware, Culture & Psychology, Vol. 13, No. 3, 307-333
The polysemic notion of `unit of analysis' has been developed by Gestalt psychologists and
fruitfully used by Vygotsky in the struggle against reductionism in psychology. Currently,
it has been appropriated by a sociocultural approach for criticizing cognitive approaches
for using the `individual as unit of analysis'. In this article, I argue that a
sociocultural approach is experiencing a crisis as it is being affected by holisma
tendency to include the universal whole in the studied phenomenon. I propose and discuss
alternatives to reductionism and holism for a sociocultural approach.
The Couple Versus the Spouse as the Unit of
Analysis in Marital Research. Revised.
Honeycutt, James M.; Norton, Robert W.- Paper presented at the Annual Meeting of the
International Communication Association (Boston, MA, May 2-5, 1982).
Abstract: A brief narrative description of the journal article, document, or resource.
Traditionally the unit of analysis in marital research has been the
individual spouse. More recently the marital relationship has often been defined as a
process of interaction and dynamic exchanges such that spouses have autonomous needs as
well as corporate needs for interdependence. Thus modern systems theory heightens the
importance of both individuals. A structural analysis of the marital dyad consists of two
types in which the aim is to determine whether relationships found at the group level are
the same or different from relationships within separate group components. The problem in
the analysis is the treatment of the couple's score so that the couple's score is not
merely a high/low spouse score in which sex effects are unaccounted for. Scores may te
summed for the husband and wife and divided by two (summation score), or scores may be
calculated to take into account the differences between each spouse's individual scores
(dispersion scores). Using both these scores as the unit of analysis in a study of marital
happiness and communicator images among 40 married couples, it is evident that the
dispersion score can reveal an effect for the degree of a couple's agreement on marital
happiness and not the level of happiness. In short, a reliance solely on individual or
couples' analyses excluded some information. Therefore, the reporting of individual and
couples' scores when a couple's analysis is used is often needed in order to give a full
picture of the couple's agreement on the criterion variable and their rank on the
criterion. However, in some situations depending on the exact research question, a
summation score or dispersion score may solely be used.
City Markets as a Unit of Analysis in Audit
Research and the Re-examination of Big 6 Market Shares - Jere R. Francis, Donald
J. Stokes, Don Anderson, Abacus, Vol 35, No 2, 1999
Abstract: Big 6 audit market shares based on aggregate national data have been used in
prior research to infer market leadership and industry expertise, and to differentiate Big
6 accounting firms from one another. In this study it is demonstrated that further
differences exist with respect to city-specific audit markets, both between firms and
within the same firm across different city markets. The specific finding is that the
national market leader is not the city-specific market leader the vast majority of time.
Usefulness of the city-level unit of analysis is further demonstrated by re-examining the
1989 mergers creating the accounting firms Ernst & Young and Deloitte Touche. The
primary effect of the Ernst & Young merger was to increase audit market shares in
cities in which the pre-merger firms already had significant market shares, resulting in
an increase in the number of cities the merged firm achieved top ranking. In contrast the
primary effect of the Deloitte Touche merger was an expansion of the number of city-level
markets in which the merged firm had significant (though not leading) audit market shares.
The findings of this study suggest that in order to move beyond our current understanding,
important audit research questions such as the reason for particular auditor-client
alignments, the competitive nature of markets, audit pricing of reputations, and auditor
reporting and independence issues should be investigated in city-level markets where audit
contracting occurs and where Big 6 market shares (and presumably reputations) vary widely
from city to city.
Units of analysis in task-analytic research
- T G Haring and C H Kennedy
Graduate School of Education, University of California, Santa Barbara 93106.
J Appl Behav Anal. 1988 Summer; 21(2): 207215. 1988.21-207. PMCID: PMC1286113
Abstract: We develop and discuss four criteria for evaluating the appropriateness of units
of analysis for task-analytic research and suggest potential alternatives to the units of
analysis currently used. Of the six solutions discussed, the most commonly used unit of
analysis in current behavior analytic work, percentage correct, meets only one of the four
criteria. Five alternative units of analysis are presented and evaluated: (a) percentage
of opportunities to perform meeting criterion, (b) trials to criteria, (c) cumulative
competent performances, (d) percentage correct with competent performance coded, and (e)
percentage correct with competent performance coded and a grid showing performance on
individual steps of the task analysis. Of the solutions evaluated, only one--percentage
correct with competent performance coded and a task analysis grid--met all four criteria.