Background: The Traditional Approach

According to Costanza and Finkelstein (Reference Costanza and Finkelstein2015), the definition and measurement of generational membership could be two major problems inherent in the literature on generational differences. So far, researchers have defined generation in terms of groups of people belonging to the same cohorts, age groups, and experience of certain common events (Joshi, Dencker, & Franz, Reference Joshi, Dencker and Franz2011). In this vein, generational difference is operationalized mostly as a categorical variable, and most researchers assign participants into generational categories based on date of birth. For instance, people born in 1958–1959 and 1973–1974 are typically divided into Boomers and Gen X, respectively (Twenge, Campbell, Hoffman, & Lance, Reference Twenge, Campbell, Hoffman and Lance2010).

We argue that a strictly age-based, categorical approach to generational difference could be problematic for two main reasons. The first problem is that a categorical approach fails to provide a clear picture for the definition of generational category. Some generational categories are made based on birth year or age. But even a straightforward classification based on birth year is done differently by different researchers, as there is no complete agreement about which birth years are associated with which categories. For example, Strauss and Howe (Reference Strauss and Howe1991) considered people born between 1943 and 1960 as Baby Boomers, whereas Lancaster and Stillman (Reference Lancaster and Stillman2002) argued for birth years between 1946 and 1964. In addition, the usage of starting and ending years is unclear and inconsistent. For example, for the Boomers generation, about which there seems to be the least disagreement on the starting and ending years, has starting years ranging from 1943 to 1946 and ending years from 1960 to 1969 (Costanza, Badger, Fraser, Severt, & Gade, Reference Costanza, Badger, Fraser, Severt and Gade2012). The birth year–based generational category may also result in false attribution to someone whose birth year is just between two different generations. For instance, someone who was born in the year 1964 may be categorized as a Baby Boomer (ranging from 1946 to 1964), although he/she may in fact have more common with Gen X (ranging from 1965 to 1981).

Another, and perhaps more important, problem is related to the mechanism through which generational differences can lead to different employee attitudes and behaviors. Factors such as common life experience are commonly used as stimuli of generational differences (e.g., Joshi, Dencker, Franz, & Martocchio, Reference Joshi, Dencker, Franz and Martocchio2010), and the existing birth year–based approach assumes that people belonging to the same generation have experienced, encoded, and integrated these stimuli in a common way (Costanza & Finkelstein, Reference Costanza and Finkelstein2015). Research in social information processing, however, has shown opposing evidence: On one hand, people can process and react to the same stimulus differently based on their different goals, values, and preferences; on the other hand, people may develop similar ways to mentally construct and make sense of different social information (e.g., Johnson, Haubl, & Keinan, Reference Johnson, Haubl and Keinan2007; Kelley, Reference Kelley1973). These findings suggest that, rather than passive receivers of what happens to them, people will actively and subjectively make sense of their life experiences. Because of various individual differences involved in this subjective sense-making process, people form different attitudes and behaviors even though they have experienced the same life event. Thus, rather than the common life experiences, people's subjective mental process of these experiences may be more effective in predicting their subsequent attitudes and behaviors. In this vein, some recent research has also argued for the importance of subjective mental experience in understanding the effects of generational difference. For example, Gardiner, Grace, and King (Reference Gardiner, Grace and King2013) have suggested that generational difference can only be effective in shaping individual behaviors when the generational stereotype converges with one's self-identity.

An Alternative Approach

Here we propose an alternative approach to solve the above-mentioned problems of generational difference in organizational research. Instead of treating generational difference as an assigned, categorical group membership, we suggest treating it as a subjective, continuous social identity. Why is this alternative approach potentially better than the original one? First, it allows the individual, rather than the researcher, to decide the generation(s) with which they identify. Second, rather than an all-or-none phenomenon traditionally used in categorical membership, a social identity approach uses a continuous scale that allows people to rate the extent (or degree) to which they identify with each category (Ashforth & Mael, Reference Ashforth and Mael1989). We believe this approach could allow researchers to study generational difference, which may explain those conflicting results in previous studies (Costanza & Finkelstein). In the following, we provide further details on the operationalization of this approach.

A subjective, continuous approach is commonly used in research on other types of social identity based on group membership. A few prominent examples can be found in research on ethnic identity (e.g., Phinney & Ong, Reference Phinney and Ong2007; Ponterotto & Park-Taylor, Reference Ponterotto and Park-Taylor2007), age identity (e.g., George, Mutran, & Pennybacker, Reference George, Mutran and Pennybacker1980; Westerhof & Barrett, Reference Westerhof and Barrett2005), and gender identity (Constantinople, Reference Constantinople1973). For instance, although gender is traditionally treated as a categorical variable of male versus female, more recent research has shown that people's social identity with certain gender groups is nevertheless a continuous mental state (Egan & Perry, Reference Egan and Perry2001), and measurements have been developed to measure gender/sex as a continuous, rather than categorical, construct (Constantinople, Reference Constantinople1973; Egan & Perry, Reference Egan and Perry2001). In addition, research on age identity provides an example of how to clarify the “gray area” on the boundaries between two different categories (e.g., the starting and ending years of a generation). For example, Westerhof and Barrett (Reference Westerhof and Barrett2005) measured age identity by subtracting one's subjective age from actual age, and they found participants’ age identities ranged from 10 (10 years older than actual age) to 35 (35 years younger than actual age).

Similar to the above-mentioned examples from other areas, we suggest a two-step approach to improve the rigor and meaningfulness of studies on generational differences. First, consistent with social identity theory (Ashforth & Mael, Reference Ashforth and Mael1989), generational difference could be defined and operationalized as the extent of people's definition and self-belongingness to different generation-based social categories (e.g., Boomers, Gen X, Gen Y). Measurement of generational identity could begin by first providing people with a list of different generation categories. This could be done through either open-ended questions or checklists that are appropriately inclusive. This method not only activates respondents’ social identity by comparing groups and thinking about relating to those groups (Tajfel & Turner, Reference Tajfel, Turner, Worchel and Austin1985) but also allows respondents to select more than one category with which they identify. Thus, this multiple-response method can be informative and potentially more accurate than the traditional method based solely on one generation category, especially in situations where a person identifies with more than one generation (e.g., someone who was born in the late 1960s may identify with both the Boomers and the Gen Xers). Moreover, this method also minimizes the inconsistency due to the use of different starting and ending years in defining generational categories because people are likely to identify with two generations that are successive in time, making it less important to achieve a definite boundary between generations based on time (year).

Second, the selected generation categories are then rated as continuous constructs. This could be measured by asking the extent/degree to which people are committed to or self-identified with a particular generation. In this way, generational identity could be analyzed as a series of continuous variables that focus on the strength of generational identity rather than just one categorical variable. This second step provides incremental value over the self-categorization step described above by providing information on the strength of identification within category. Moreover, as continuous variables can provide better prediction compared with discrete categorical variables, it is reasonable that studying generational difference as a series of continuous variables may also provide stronger and more consistent effects on work attitudes and behaviors.

Here we present a hypothetical example of how generational difference can be operationalized as a series of continuous variables in actual organizational research. A group of researchers is trying to study how nurses’ generational differences can influence their interpersonal integration, as previous research has shown that intergenerational conflict is an important occupational stressor for the nurses (Santos & Cox, Reference Santos and Cox2000). The researchers decide to operationalize generational differences as the difference in the strength of generational identity with each generational group. To measure generational identity, the survey includes a list of labels for different generational groups (e.g., Baby Boomers, Generation X, and Generation Y). First, participants are asked to select one or more generational labels with which they would like to identify themselves. They are then asked to evaluate the extent to which they agree that they have a strong sense of belonging to the chosen group on a Likert scale (1 = slightly agree, 2 = somewhat agree, 3 = mostly agree, 4 = totally agree). The score on each generational group can range from 0 (if a participant does not choose this group) to 4, which represents a participant's strength of self-identity with this generation. In this case, the larger the score difference between two participants on each generational group, the more they differ in generational identity with this group and the more likely there exists intergenerational conflict.

Concluding Thoughts

In sum, we argue that a subjective, continuous approach would help researchers to better understand generational effects at work. By treating generational difference as a series of continuous variables rather than just one categorical construct, this approach provides a more theoretical, nuanced definition of generational differences, which should allow a more reliable valid measurement. Improving conceptualization and measurement of generational differences should result in a stronger prediction of their effect on work attitudes and behaviors.