[Main page]

Methodological Triangulation,
Or How To Get Lost Without Being Found Out
by Alexander Massey BA PGCE MA MSc

Reproduced by kind permission of JAI Press
from Massey, A. and Walford, G. (Eds.) (1999) Explorations in methodology,
Studies in Educational Ethnography, Vol.2, Stamford, JAI Press, 183-197


1. STARTING THE JOURNEY

"The English language is an arsenal of weapons. If you are going to brandish them without checking to see whether or not they are loaded, you must expect to have them explode in your face from time to time." (Stephen Fry, The Liar)
1.1 Looking ahead: aims of the chapter

In his book 'Images of Organisation', Morgan (1986) shows how metaphor can illuminate and challenge our everyday thinking, but also constrain it, leading sometimes to the uncritical adoption of certain views and the ill-advised neglect of others. He argues that "the use of metaphor implies a way of thinking and a way of seeing that pervade how we understand our world generally (p.12)." In particular, he suggests that "metaphor exerts a formative influence on science (p.13)."

The purpose of this chapter is to examine the metaphor of triangulation - in particular methodological triangulation - in the social sciences, to demonstrate the power it still holds over some researchers, and to argue that trust in this form of triangulation is misplaced. Blaikie (1991:131), in a very clear and convincing article, argues that 'triangulation means many things to many people and ... none of the uses in sociology bears any resemblance to its use in surveying.' However, this chapter will show that some researchers have mistakenly assumed that the ontological and epistemological bases of certain sociological activities are the same as those underpinning the triangulation methods used in surveying.

The result of this philosophical and methodological confusion is that in studies that use multiple methods, many misleading and invalid claims are made in the name of triangulation. This has profound implications for ethnography, since one of its defining characteristics is that it uses multiple methods (Massey and Walford 1998:6).

In order to head off some of the confusion which can arise from doing multiple method research such as ethnography this chapter has two parallel objectives. The first is to highlight an area of philosophical dispute, and the second is to identify a number of common logical errors underpinning the practice of so-called methodological triangulation (Table 1).

Table 1. Common Error Types in Methodological Triangulation

Type A
Using a second method to 'prove' the truth of a first method, rather than simply define it as true.
Type B
Claiming that agreement between the results of two methods 'proves' the validity of the second method as well as the first (the principle of mutual confirmation, also known as 'arguing in a circle').
Type C
Taking answers that look the same to mean the same thing.
Type D
Assuming that the researcher can accurately convert a qualitative statement by a respondent in such a way as to plot it on the same place in a scale as a respondent would if asked (a development from Error Type C).
Type E
Assuming that propositions and answers derived from different methods can converge or diverge (i.e. 'agree' or 'disagree').
Type F
Believing that the 'strengths' of one method can offset the 'weaknesses' of another, leading for some researchers to the illusion of a problem of how to prioritise the findings resulting from different data sources (the 'weighting problem').
Type G
Comparing the results of two samples as though they belong to the same population when there is no methodological or statistical demonstration that they do.

1.2 Where the journey started: the origin of triangulation

Blaikie (1991), a land surveyor for 16 years before turning sociologist, gives a clear and detailed description of the original concept of triangulation in surveying, navigation and military strategy. Essentially, the aim is to establish the position of a point, which can be achieved several ways:

  1. A point can be located 'from two others of known distance apart, given the angles of the triangle formed by the three points (Clark 1951:145).' The absolute position of the point can be established only if the absolute positions of the two reference points are known. Otherwise, the most one can establish are the positions of the three points relative to each other.

  2. A topographical feature can be plotted by observing it 'from a number of known positions, thus forming a triangle in which one side and the adjacent angles are known (Blaikie 1991:118).' In this method (intersection), the absolute as well as the relative position of the feature can be calculated.

  3. An unknown position can be fixed by measuring, from it, the angles to at least two other known positions (resection).

The ontological foundation of each of these three forms of triangulation is that there is a reality separate from the observer. The epistemological assumption is that the set positions are not open to interpretation, but can be established through a direct correspondence between the positions and sensory experience of them.

1.3 'This path looks promising!' The appropriation of triangulation by the social sciences

Knafl and Breitmayer (1989:227) suggest that the appropriation of the concepts of triangulation preceded the adoption of the term itself in social science research:

It was first used metaphorically in the social sciences to characterise the use of multiple methods to measure a single construct (Campbell 1956; Campbell and Fiske 1959; Garner 1954; Garner, Hake and Eriksen 1956), a practice also referred to as multiple operationism, convergent operationism, operational delineation, and convergent validation (Campbell and Fiske 1959).

In the literature on triangulation in the social sciences, it is usually Webb et al. (1966) who are attributed with the first use of the term itself. This early thinking and writing was soon to be taken up enthusiastically in research methods textbooks (Denzin 1970; Smith 1975), thus reinforcing the use of triangulation as a legitimate technique within the social sciences research which has continued to this day (e.g. Hammersley and Atkinson 1995). It is not difficult to find numerous references to triangulation throughout methodological writing in social research. For example, Mathison (1988:15) states: 'Extending this metaphor to social phenomena, the researcher (navigator or military strategist) can use several kinds of information to determine the truth (location) about some social phenomenon (an island or the enemy).' Moreover, triangulation is a standard topic on methodology courses where, while seen as sometimes problematic, it is nevertheless regarded as appropriate in social scientific thought.

Typologies of triangulation have proliferated: data, investigator, theoretical, methodological, multiple, between-methods and within-methods triangulation (Denzin 1970; Jick 1983); simultaneous and sequential triangulation (Morse 1991); planned and unplanned triangulation (Deacon et al. 1998). Knafl and Breitmayer (1989) have managed to organise these many types into two strands based on what they see as the two principal aims of triangulation in the social sciences: convergence and completeness.

In the measurement of discrete variables, triangulation contributes to the investigator's efforts to achieve confirmation or convergent validity. In studies that address more encompassing domains of interest, multiple triangulation contributes to the investigator's ability to achieve a complete understanding of that domain (Knafl and Breitmayer 1989:237).

Both of these aims of triangulation will now be examined in detail in order to show that their conceptual basis is fundamentally flawed to such an extent that generations of researchers and readers have lost their way through their very attempts to improve sociological 'navigation' techniques.

NEXT page (section 2) >>>>>>