The Nominal Group Technique: Its Potential for Consumer Research

JOHN D. CLAXTON

J. R. BRENT RITCHIE

JUDY ZAICHKOWSKY*

This article illustrates the adaptation and modification of the Nominal Group Technique from the field of organizational planning for use in the study of consumer behavior. The method is compared to traditional group interviews and structured survey methods.

The purpose of this paper is to introduce consumer researchers to a recently developed group interview technique that provides an attractive alternative to existing research methods. This procedure, termed Nominal Group Technique (NGT), was originally developed by Delbecq, Van de Ven, and Gustafson (1975) as an organizational planning technique. Although the developers indicated that NGT may be useful in consumer research (Delbecq and Van de Ven 1971; Delbecq, Van de Ven, and Gustafson 1975), few such applications could be found.

The NGT method offers two major advantages. First, although a group technique, the structured nature of the session output makes the analysis of individual perceptions possible. Second, using the data analysis extensions developed in the current research, it is possible to do both intra and intergroup comparisons.

The discussion that follows is divided into four major sections. The first section describes NGT as it is normally employed. This general discussion is followed by a specific example of how the NGT approach has been applied in consumer research. The third section focuses on issues of data analysis peculiar to this type of consumer research, and, finally, a summary is presented of the advantages and limitations of NGT in comparison with alternative data collection methods.

* John D. Claxton is Associate Professor of Marketing, University of British Columbia, Vancouver. Canada V6T I W5. J. R. Brent Ritchie is Professor of Marketing, University of Calgary, Alberta, Canada T2N IN4. Judy Zaichkowsky is a doctoral candidate at University of California at Los Angeles, Los Angeles, CA90024. The authors express their appreciation to the Consumer Research and Evaluation Branch of the Department of Consumer and Corporate Affairs, Government of Canada, for the financial support that made this study possible. They are also grateful to Pierre Filiatrault, Universite du Quebec a Montreal, for his contribution to the research.

NOMINAL GROUP TECHNIQUE

The NGT Process

The NGT procedure, as used for program planning, is normally implemented in six stages. Participants are first presented by the session moderator with an initial statement of the topic area to be discussed. For example, top management of a corporation might be asked to indicate what directions future diversification would take. Once it is clear that participants understand the issue, further discussion is halted.

Participants are then directed to reflect individually on the topic and to record their personal responses on a worksheet containing a written statement of the issue being addressed. This period of individual reflection and recording of responses usually lasts from five to 20 minutes, depending on the complexity of the topic under discussion.

The group moderator subsequently asks a participant, chosen at random, to state one of the responses s/he has arrived at individually. This response is written in a concise yet complete manner on a large flipchart. At this point, the participant is allowed to explain his/her response briefly, so that its meaning is clear to other participants. This process is repeated in roundrobin fashion until all participants have had a chance to express a response. Second and thirdroundsmay follow, depending on the number of ideas identified by members. Participants are allowed, and even encouraged, to express additional ideas that have been stimulated by the remarks of others.

The next stage involves consolidation and review, of the complete set of ideas. At this point all flipchart sheets are posted so that all responses are visible. The moderator reviews the responses recorded on the flipcharts to eliminate duplications and to ensure that all responses are clearly understood by participants. Each response is then assigned an identifier code, such as a letter of the alphabet.

Participants are subsequently requested to establish the relative importance that should be accorded to each of the response ideas. This importance may reflect, for example, the desirability of a given idea for corporate diversification. Although various approaches may be employed to establish the importance of each response (Delbecq et al. 1975), a commonly used method is to instruct each individual first to select a certain number of responses (e.g., eight)' that he considers to be most important. The participant then writes each of these responses on a 3 x 5 card along with the alphabetic identifier, and is asked to rank the eight responses in terms of their relative order of importance.

The final stage is compilation of the results. In this stage. the rankings accorded to the various ideas by each participant are aggregated to provide a measure of overall importance. As per the Delphi technique (Linstone and Turoff 1975), these results may be presented to participants and a second round of ranking undertaken to permit individuals to adjust their judgements in the light of the earlier evaluations. However, this round is not essential unless the initial judgements were highly variable or the purpose is to achieve a reasonable level of group consensus.

To summarize, the NGT process is a systematic approach designed to provide two specific types of output. First, it provides a list of ideas relevant to the topic in question. Second, the technique provides quantified individual and aggregate measures of the relative desirability of the ideas raised in the session.

APPLICATION OF NGT TO CONSUMER

RESEARCH

Nominal Group Technique, as discussed by Delbecq et al. (1975), has been applied primarily to group decision making (problemidentification. problemsolving processes) in two general types of situationsintraorganizational group decision making and soliciting expert or citizen views as input for public policy formulation.

The specific example that follows is drawn from a study that examined the potential of NGT for consumer behavior. The overall purpose was to assess consumer shopping problems and prepurchase searchtradeoffs associated with the selection of consumer products and services. Nominal Group Technique was used to identify consumer perceptions of major prepurchase search problems.

The NGT sessions covered five product/service categories. The following discussion of NGT focuses on one of these, automobile repair services. The purpose is to identify the relative merits of NGT by discussing the advantages of the technique and the problems encountered during its application.

Method

Twenty (20) NGT sessions involving 224 participants were carried out in five metropolitan areas (four sessions per city). These cities included one metropolitan area from each of the five major geographic regions in Canada. The four sessions conducted in each city were segmented, as follows:

  • Two sessions were held in upper middle class neighborhoods; two in lower middle class neighborhoods.
  • In each neighborhood, one session was conducted with male participants and the other with female participants.
  • The sessions with male participants considered consumer shopping problems on:
  • automobile repair services,
  • home repair and renovation services, and
  • furniture and appliances.
  • The sessions with female participants considered consumer shopping problems on:
  • grocery products,
  • clothing and footwear, and
  • furniture and appliances.

Nominal Group Technique Example: Automobile Repair Services

Each NGT session dealing with automobile repairs lasted approximately 30 minutes. The total interview period, lasting from two to three hours, was composed of two major elements:

1. NGT sessions involving each of three different products/services

2. Administration of structured questionnaires concerning shopping problems, actions to overcome shopping problems, and demographics.

As per the general discussion of NGT presented earlier,the discussion of automobile repair problems was initiated by providing participants with a worksheet stating the issue to be addressed. Using this sheet as a guide, each participant spent 10 to 15 minutes identifying the shopping problems he had encountered with respect to automobile repair services.

These problems were then elicited from participants on a roundrobin basis, and written in order of presentation on sheets of a flipchart that were posted to be easily visible. Once all major ideas had been exhausted and the responses reviewed, each group member selected the eight problems from the complete list that he considered most serious. He then wrote these eight problems on 3 x 5 cards, and ranked them frommost to least serious.

ANALYSIS OF NGT DATA

Each NGT session concerning automobile repair problems provided two forms of raw data:

  • A list of all problems identified by the group. On average, this represented 17 problems per session.
  • Each participant's rank order of importance for theeight problems he considered the most serious.

The data resulting from these ten sessions consisted of 170 problem statements. An important characteristic of this data bank of problem statements is that the statements from different sessions were not directly comparable in their raw state because the same issues were not necessarily raised in all sessions, or were not put forward with exactly the same wording. As a result, there was a need to simplify the structure of the many qualitative statements into a more easily understood form that would permit comparison across NGT sessions. Because this particular data analysis problem was not present in prior applications that employed only a single NGT group, it was necessary to develop a standardized method of analysis that allowed the technique to be extended to multiple NGT session situations. The method developed consisted of four basic steps:

1. Categorization of initial problem statements intoproblem themes

2. Calculation of a score or index reflecting the importance of each problem theme

3.Ranking of problem themes according to their importance index

4. Regrouping of problem themes to form major problem dimensions.

Identifying Problem Themes

The purpose of the first step was to look for themes common across NGT sessions. The procedure used was to prepare individual cards for each statement. All statements (from all sessions) were then categorized independently by each of the three researchers. These three categorizations were compared for consistency through content analysis procedures. Each problem category, or problem theme, involved statements containing essentially the same words or ideas. As might be anticipated, the number of themes established varied with the type of product/service in question (automobile repair services 24, clothing/footwear 32, and

EXHIBIT

MAJOR PROBLEM DIMENSIONS IDENTIFIED FROM

PROBLEM THEMES: AUTOMOBILE REPAIR SERVICES

Dimension I: Obtaining time/cost commitments

Theme 15Difficult to arrange appointment/servicing of car

Theme 16Repairs not done when promised

Theme 12Estimates hard to obtain or grossly inaccurate

Dimension II: Incompetent workmanship

Theme 6Incomplete servicing/repeat visits required

Theme 20Unbelievable incompetence of mechanics

Theme 7Difficult to locate competent mechanics or reputable firms

Dimension III: Problems with guarantees

Theme 5Lack of guarantee on work

Theme 9Poor servicing under warranty

Theme 10Warranty is inadequate/warranty information is misleading

Theme 11Extra costs arising from warranty work

Dimension IV: Lack of integrity

Theme 1Garage owner's lack of integrity by taking advantage of consumers through unfair practices, so they can make more money

Theme 2Not sure that the time charged represents the real time spent on repairs

Theme 18Have to purchase a complete unit when all you need is a small part

Theme 21Work carried out by apprentice, but billed at rate of specialist

Theme 13Planned obsolescence/rust

Dimension V: Lack of personal concern for customer

Theme 3Depersonalized serviceno contact between car owners and mechanic

Theme 4Not having consumers interest at heart, lack of professionalism

Theme 17Unavailability and long delivery time for parts

Dimension VI: Repair cost problems

Theme 23Cost of repairs much too high

Theme 22Wide variation in cost across garages (even oil changes)

Theme 24High cost and difficulty in getting foreign cars serviced

Theme 19Difficult to judge value for money when buying parts

Dimension VII: Industry conditions

Theme 8Too many selfserve stations and not enough garages and mechanics to go around

Theme 14Difficult to do repairs due to metric system/electronic design/government standards

furniture/appliances 24). As an example, the actual themes identified for automobile repair services are presented in the Exhibit.

Index of Theme Importance

Step two involved the calculation of an index of theme importance. The index employed was designed to reflect two different measures of importance, namely, the frequency with which a problem was selected by respondents as important, and the priority accorded to it when it was selected. Thus, a theme that had beenidentified in most sessions and that had been ranked highly by most participants would receive a high index score; one selected frequently but considered less important, or selected less frequently but judged important by certain participants, would receive an intermediate index score; and a theme selected infrequently and accorded little importance would receive a low index score.

In the present study, a participant's score for a particular theme was based on the rank importance accorded by that participant. Statements ranked highest in importance received a score of 8, items second in importance received a score of 7, and so on. Theme scores were then aggregated across all participants to provide a summary index of theme importance.

Ranking Themes and Identifying Problem

Dimensions

The index of theme importance was used to rank the different problem themes within a product/service category in terms of their approximate order of importance. For many studies, this ranking of themes may represent sufficient analysis, as it provides the researcher with a quantitative measure of the importance of the various ideas expressed during the NGT sessions. In the present study, the intent was to further aggregate the problem themes to identify what were termed major problem dimensions. These dimensions were subsequently employed as the basis for analyzing consumer retail service tradeoffs through conjoint measurement (Green and Wind 1973). The manner in which the 24 auto repair service themes were regrouped into problem dimensions provides one specific example of this aggregation process (Exhibit).

Although aggregation of themes is useful in certain instances, it results in a much simplified version of the original data and provides less understanding of the original problem statements. As such, aggregation of themes into dimensions should not be done with a view to subsequently forgetting their existence, as the greatest insight into the understanding of consumer behavior appears to be gained at the theme level.

It is also useful to note the difference between identifying themes and identifying dimensions. The purpose of the former is to aggregate across NGT sessions statements that express essentially the same idea, a process conceptually similar to content analysis (Holsti 1969; Kassarjian 1977). On the other hand, identification of dimensions is done for the purpose of providing a typology of themes, i.e., grouping themes that relate to a general problem area.

DISCUSSION

What are the major strengths and weaknesses of NGT for consumer research, and where does NGT fit relative to other data collection procedures? Severalstrengths can be identified. First, although a group method, NGT provides structured output that can be analyzed at an individual level. The term Nominal Group Technique is intended to suggest that the method is nominally viewed as group based, though for the most part the activities and output focus on individual efforts. The early stages of the process provide respondents with the opportunity to hear the views of others as they are thinking through the topicunder discussionsimilar to other group methods. On the other hand, the final stages require respondents to sort and rank the items generated in earlier stages. Thus, the data output is more structured than is usual with group methods.

Second, the NGT process results in high respondent involvement and commitment. This commitment develops as respondents express their views to others in the group and realize that they are sharing in the identification of items to be evaluated. This advantage can be particularly useful when participants need time to think through their responses. This is not to imply that they are thinking up new ideas in a creative sense. Rather, when asked to recall behavioral experience, it is probable that respondents have had little experience identifying the various steps that were involved in the behavior. In other words, although the behavior of interest may be current or even habitual, the process of articulation requires time and commitment to recall the various components.

Third, the process of identifying and scoring problem themes, as developed in this application, makes it possible to study both intraand intergroup differences.

The major disadvantage of NGT relates to sampling. Because participants have to agree to come to a central meeting location, attempts at probability sampling are met by a serious level of nonresponse, as discussed further in the following sections.

Comparison with Focus Group Interviews

A number of authors provide reviews of qualitative research methods (Bellenger, Bernhardt, and Goldstucker 1976; Bogdan and Taylor 1975; Higginbotham and Cox 1979). One of the most insightful of these reviews has been presented by Calder (1977) in which he distinguishes among the exploratory, clinical, and phenomenological dimensions of qualitative research. Although NGT possesses some characteristics of each of these dimensions, it is perhaps closest to what Calder defines as exploratory research. As he points out, a major strength of exploratory methods is the ability to identify major issues or attributes associated with a particular research problem. However, there are several characteristics of standard focus groups that restrict this method to exploratory applications. First, the output of the session is relatively unstructured. Although the session can provide an extensive list of attributes, the process does not facilitate establishing attribute priorities. Second, a small subset of the participants may be outspoken and dominate or intimidate the rest of the group. Third, to minimize this potential domination and to ensure the desired depth of coverage, a highly trained session leader is required. Finally, the necessity to bring participants to a meeting room virtually precludes the use of probability sampling procedures. However, the quota sampling methods usually adopted are entirely consistent with the exploratory nature of the focus group process.