Step 1: Identify Causal Statements

The first task is to identify the causal statements from the documents (e.g., interview

transcripts or annual reports) (Axelrod, 1976). This process involves identifying the

cause and effect phrases and the linkage between them. Causal statements are statements

that imply a cause-effect relationship. Some of the key words used in identifying explicit

causal statements are “if-then,” “because,” and “so” (Axelrod, 1976). In addition to

explicit causal statements, according to Axelrod (1976), there are also implicit relationships

found in causal statements. The phrase may not contain the traditional key words

used to identify causal statements, but the causality of the sentence is clear within the

context of the text. Some “key words” that have been used in identifying implicit causal

statements are “think,” “know,” “use,” and “believe”. For example, the sentence “If I

want to get beyond where I am today, then am I going to have to go outside of the

business?” could be coded as an explicit statement since it contains the words “if” and

“then.” Additionally, the sentence “I don’t think gender should be an issue, I would

promote whoever is smartest” can be coded as an implicit statement. The statements in

the form of concepts and cause-effect relationships are captured in the language of the

Term Definition

Causal Map A network of causal assertions (cause/link/effect) that can be

expressed in a matrix or diagram form.

Causal Statement A statement (phrase or sentence) that contains a casual assertion, most

generally of the form cause/link/effect.

Coding Scheme A dictionary of terms (concepts or constructs) and definitions of those

terms (concepts or constructs). The coding scheme is used to simplify

the causal statements and corresponding maps.

Concept A word or phrase that captures the meaning or essence of a

participant’s phrase.

Construct A word or phrase that captures the meaning or essence of a group of

concepts.

Link The relationship or causal belief between two concepts (or constructs).

Raw Causal Map A causal map in which the concepts (constructs) are represented in the

language of the participant.

Raw Causal Statement A causal statement that is captured in the language of the participant.

Revealed Causal Map The assertions of causality the participant chooses to reveal to the

world.

Table 3. Causal mapping definitions

Figure 2. Revealed causal mapping process

participants (Narayanan & Fahey, 1990). Other examples of causal statements would

include:

1. Object-oriented development is easy because you think of everything as an object.

2. If I’ve got this object built up then I go back and actually try to write some of the

methods.

3. Once I have all of the information I need I think about what are the objects that will

be needed.

Depending on the type of data collection, IECM or TBCM, the coding process will differ.

If you are using TBCM, generally you are using public documents (e.g., annual reports),

which have been carefully crafted. The author of the document has (most likely) placed

emphasis on the sentence construction, grammar and intended meaning of each sentence.

In this context, the causal statements should be relatively clear and straightforward.

*The term “text” is used to represent both IECM transcripts and TBCM texts.

In contrast, if you are using IECM, the causal statements are often difficult to discern

(Kemmerer, Buche & Narayanan, 2001). In this case the participant sample plays a large

role in the ease (or difficulty) of coding. For example, if you are speaking to IS personnel

regarding their current project, they are usually quite articulate. In contrast, if your

research sample consists of IS students discussing a very technical topic, or respondents

discussing a sensitive topic (e.g., layoffs) the participants may have difficulty

expressing themselves. In addition, you will probably have several “starts and stops”

within the transcript. By this I mean an individual will begin to speak, stop and then restart

with the thought. This can present challenges when coding the transcript. In this case,

it is up to the researcher to discern the causal statement (if any) in the text. It is often

helpful to have an audio recording (if possible) to listen to the tone of the participant in

addition to the words.

Identification Rules

The guidelines, which have been adapted from Axelrod (1976), are provided to show

researchers how causal maps can be derived from texts. The coder must scrutinize the

text to record all cause-effect relationships within the text. The sentences or phrases that

are of interest to the coder are those that assert a causal relationship (A affects B). To

appropriately identify the causal statements the researcher needs a set of decision rules

to help guide the process. The rules are:

1. Some relationships are implicit in the phrase or sentence and a cause/effect

relationship cannot be found in the structure of the phrase. In this case the coder

should ask herself if the phrase implies a relationship between variables. If yes, then

the phrase should be coded as a causal statement (be careful not to insert bias into

coding implicit statements to create assertions).

2. It is important to maintain the original language of the participants as faithfully as

possible.

3. It is important to reflect the speaker’s statement in kind and number. If a speaker

states a relationship more than once, the coder should note the relationship each

time it is mentioned.

4. If a speaker agrees with an assertion made by someone else the coder should pay

close attention to the speaker’s wording. If the speaker is agreeing with the

assertion then it is recorded as a causal statement. If the speaker is merely

acknowledging the statement then it is not coded.

5. Assertions should be made within a sentence or two at most. Do not look for

assertions by linking paragraphs.

In addition to these basic guidelines, Wrightson (1976) has provided a listing of the

structural relationships that may be found within a text and how they should be coded.

See Appendix A for an adapted (and abbreviated) sample of these structures.

Table 4. Sample causal statements