Intuitive and Instrumental Grief: A Study of the Reliability and Validity of the Grief Pattern Inventory

Abstract

A convenience sample of 209 participants completed the Grief Pattern Inventory (GPI) together with the Hogan Grief Reaction Checklist (HGRC), the Integration of Stressful Life Events Scale (ISLES), and the Myers–Briggs Type Indicator (MBTI). Alpha coefficients of the GPI’s intuitive and instrumental subscales were improved by eliminating low-performing items and empirically reassigning items from the GPI dissonant subscale. The two modified scales showed a near zero intercorrelation indicating they were independent constructs in this sample, a conclusion further verified by factor analysis. Both styles correlated with distress measures from the HGRC and correlated negatively with adaptation indicators from the ISLES, though the findings were stronger for intuitive grievers. MBTI findings revealed that intuitive grievers endorsed significantly more “feeling” in how processing occurs while grievers who preferred sensing/thinking functions more often identified with the instrumental grief pattern. Implications of these findings for scholars and clinicians are discussed.

Keywords

attachment gender mourning grief death

Many clinicians performing grief therapy and bereavement care have adopted Doka and Martin’s (2010) theory of three distinct grieving styles found among the bereft: intuitive, instrumental, and dissonant. Doka and Martin argue that recognizing an individual’s preferred grief pattern can enhance clinicians’ sensitivity to variations in personal expression and lead to better understanding of the individual griever. Knowledge of grieving styles is also thought to enable clinicians to devise more customized, targeted intervention strategies, a notion consonant with the contemporary emphasis on empirically based therapies.

According to Doka and Martin’s (2010) nosology, intuitive grievers respond affectively to their loss and often experience a range of painful emotions. Intuitive grievers gain strength and solace from sharing their experiences with others. They benefit from exploring and coming to understand their interior emotional landscape, openly expressing or ventilating negative emotions, and receiving support and encouragement from others, that is, “talking it out.” In psychotherapy where patients come for assistance with grief and bereavement, individuals displaying the intuitive pattern are commonly represented.

For Doka and Martin’s (2010) instrumental grievers, there is a greater focus on cognitions while painful negative affect is moderated. Such individuals are not without feelings; rather they prefer to maintain mastery and control over their emotions and to “think through” the conundrum of grief, if at all possible. Their complaints of discomfort are less likely to be emotion based and more likely to be somatic in nature. Instrumental grievers are less likely to seek psychotherapeutic help but, when they do, they often seek information, coping strategies, and practical suggestions about how to manage the experience of losing someone close to them. Instead of talking out their grief, they would rather “work it out.”

Grievers showing the dissonant pattern are thought to be generally maladaptive by virtue of being “at war with themselves” (Doka and Martin, 2010, p. 91) about how they handle the loss of a loved one. For example, intuitive grievers may suppress emotion in an unnatural manner and try to ignore the internal disharmony this suppression creates. Or, they may worry more about image management, that is, being seen as “in control,” so they suppress emotion. Instrumental grievers may feel guilty or disingenuous about a lack of emotion and so question themselves. Both types may engage in self-numbing with alcohol or drugs. These short-sighted strategies may result in nonresolution long term.

As conceptually appealing as Doka and Martin’s (2010) model may be, there has been relatively little empirical validation provided for their theory of grieving patterns. Unfortunately, this situation runs the risk of clinical practitioners adopting theoretical constructs, and making extrapolations from those constructs, without really having an empirical basis for the assertions. In the contemporary era where empirically supported treatments are considered the standard of care, grief professionals relying on unverified constructs risk the legitimacy of their work. The current study is an effort to redress this dearth of quantitative evidence regarding Doka and Martin’s theory of grief patterns.

Literature Review

In their first book tracing the evolution of their original ideas about masculine and feminine grief, Martin and Doka (2000) proposed a noncategorical view wherein they conceptualized grief expressions along a continuum between intuitive and instrumental grief patterns, with most people thought to display a blend of these two styles. They posited that an individual’s grief pattern was influenced not just by gender but by a whole array of biopsychosocial influences including factors such as culture, socialization, birth order, and temperament.

Intending to further explore the idea that grief patterns manifest along a continuum of styles, Martin and Wang (2006) tested a pilot version of the Grief Pattern Inventory (GPI), a 14-item scale designed to differentiate intuitive and instrumental grief patterns. They studied 63 individuals who were part of a Mid-Atlantic college community, including students, faculty, and staff. Their sample included 43 women and 20 men with an average age of 31 years. Country of origin and religious background were varied and no specific information was provided about the sample regarding these variables. Unfortunately, Martin and Wang did not report alpha coefficients for the intuitive and instrumental subscales, an omission precluding any psychometric appraisal of subscale reliability, although they did report moderate intercorrelations among items on the respective subscales. They also found a significant negative correlation between the intuitive and the instrumental subscales (r = −.525) lending some credence to the idea that these styles are opposite in nature. Women scored significantly higher than men on the intuitive subscale; men tended to score higher on the instrumental subscale, but not to a statistically significant degree. An attempt at establishing construct validity was made by comparing the scores of instrumental and intuitive grievers on the Grief Experience Inventory (GEI; Sanders, Mauger, & Strong, 1985). The primary finding from this effort showed that intuitive grievers scored higher on despair as well as loss of control while instrumental grievers scored lower on loss of control.

Along with publication of their second book, Doka and Martin (2010) provided an expanded version of the GPI, now a 30-item scale with three subscales: intuitive, instrumental, and dissonant. A scoring system was provided but no reliability or validity statistics for the 30-item GPI were reported. The availability of this expanded scale created an opportunity for researchers and scientists to empirically investigate the concepts of intuitive and instrumental grief and their potential utility to inform the practice of grief therapy.

In an unpublished doctoral dissertation on factors influencing adjustment following romantic break-ups, Primeau (2013) used the 30-item GPI when assessing 356 Midwestern college students. Inclusion criteria required participants to be unmarried, 18 to 25 years old, U.S. domestic, undergraduate, heterosexual, and have a history of a romantic breakup in the past 2 years (the index loss for her study). Her sample was 62% women and 87% White. Although no alpha coefficients to assess reliability were reported, Primeau found a negative correlation between the intuitive and instrumental subscales (r = −.38). Furthermore, her data revealed that men scored significantly higher than women on instrumental grief but no gender differences were observed on the intuitive subscale. Students higher on intuitive grief scored lower on overall social adjustment and personal emotional adjustment.

An interesting finding pertaining to attachment style emerged from Primeau’s (2013) results. After a romantic breakup, students scoring higher on intuitive grief reported significantly higher scores on a measure of attachment anxiety while students higher on instrumental grief reported higher scores on attachment avoidance. From a construct validity perspective, these latter findings suggest that intuitive grievers may consciously carry more subjective distress about their loss while instrumental grievers are more likely to distance themselves from emotional expression regarding their loss.

In a recent study of attachment dynamics between male dog owners and their canine companions, Bartone and Blazina (2016) also used the 30-item GPI as one of their psychometric measures. A total of 911 adult men were recruited with a sample composition of 72% between ages 18 to 44, 71% White, 52% married/partnered, and 48% Christian. Of them, 551 completed the GPI which yielded Cronbach’s alpha coefficients for the intuitive (r = .78), instrumental (r = .78), and dissonant (r = .72) subscales indicating acceptable scale reliability. The intuitive and instrumental subscales did not significantly correlate with each other (r = .118), but both intuitive and instrumental were significantly and positively correlated with the dissonant subscale (r = .299 and r = .576, respectively). In addition, the dissonant grief pattern was more prominently represented among younger men.

Within Bartone and Blazina’s (2016) sample, 194 of the men reported loss of their canine companion, the majority (68%) having experienced death of their pet between 3 and 5 years prior to the study. The investigators found significant correlations between anxious attachment style and intuitive (r = .286) as well as dissonant (r = .259) subscale scores. Significant correlations were also found between avoidant attachment style and the instrumental (r = .372) and dissonant (r = .433) subscales. These findings certainly suggest that dissonant grievers have a much higher probability of displaying an insecure attachment pattern. Bartone and Blazina’s data also corroborates Primeau’s (2013) findings of an apparent association between intuitive grieving/anxious attachment and between instrumental grieving/avoidant attachment.

As intriguing as these results may be for professionals eager to apply the concept of grieving patterns to clinical work, it is imperative to recognize that neither of these studies (Bartone & Blazina, 2016; Primeau, 2013) investigated a sample of individuals mourning the death of a significant person in their lives, the actual life event on which the concepts of intuitive and instrumental grief are predicated. While these data begin to provide some answers about the reliability and validity of the GPI, essential questions remain. Specifically, are the intuitive, instrumental, and dissonant subscales of the GPI demonstrably reliable? What validity indicators beyond attachment style empirically anchor the constructs of intuitive and instrumental grief?

Method

The data used in this investigation were drawn from a larger set of information collected by the third author (LP-D) for her dissertation on the relationship between grief and personality (Prosser-Dodds, 2013).

Participants

A convenience sample of 271 participants was initially recruited by e-mail invitation from three venues around the United States: a large memorial chapel in a Midwest metropolitan area; a faith community in the Northeast United States; and seminars on Emotional Intelligence given around the United States by the third author (LP-D). Participants met three inclusion criteria: bereaved because of the death of a loved one more than 6 months prior to the study; over the age of 25; and able to read and comprehend English at a sixth grade level. Among the sample, 62 respondents lacked complete data because of tabulation errors in the online computer program which failed to record some of the responses on two different questionnaires. Thus, the final sample for this investigation consisted of 209 participants.

The sample was almost exclusively White (94.7%), majority women (69.4%), and represented the following age cohorts (in years): 24–30 (n = 21; 10.0%); 31–40 (n = 30; 14.4%); 41–50 (n = 62; 29.7%); 51–60 (n = 62; 29.7%); 61–70 (n = 29; 13.9%); and 71 + (n = 4; 1.9%). A variety of relationships between participant and decedent were represented with the deceased person identified in the following frequencies: spouse/partner (n = 12; 6%), parent (n = 19; 9%); child (n = 86; 41%); sibling (n = 17; 8%); other relative (n = 59; 28%); friend (n = 14; 7%); multiple losses including parent and siblings (n = 2; 1%). The average length of time since the death was 7.01 years (standard deviation = 8.19) with a range of 6 months to 44 years. Of the 209 participants, 23 (11%) had sought some form of grief counseling in response to their loss.

Instruments

The GPI (Martin & Doka, 2010) is a 30-item scale with three subscales: intuitive (11 items); instrumental (12 items); and dissonant (7 items). This instrument uses a Likert-style rating of 30 loss-related statements to determine the dominant grief style of the bereaved individual. Summary scores for each of the three proposed subscales can be derived according to the authors’ delineation of which items comprise each of the three theoretical subscales. Despite its conceptual popularity, basic reliability and validity information in samples of bereaved persons has not yet been published. That dearth of evidence was precisely the reason for the current investigation.

The Hogan Grief Reaction Checklist (HGRC; Hogan, Greenfield, & Schmidt, 2001) is a 61-item self-report, Likert-style questionnaire comprising items empirically generated from written and recorded statements made by bereft individuals. The items sorted into five factor-analytically derived scales of grief misery: despair, panic behavior, blame/anger, detachment, and disorganization. In addition, a sixth scale measures personal growth following bereavement. Reliability was demonstrated by Cronbach’s alpha coefficients that range from .79 to .90 for the six subscales (Hogan et al., 2001), findings replicated by Gamino, Sewell, and Easterling (2000) who recorded reliability coefficients from .71 to .86. Previously, we found the HGRC to be more sensitive than another widely used grief measure, the Grief Experience Inventory (GEI; Sanders et al., 1985), in detecting subtle differences in grievers’ self-reported negative affect (Gamino et al., 2000). The five “misery” subscales of the HGRC can be summed to provide an overall grief misery score. Scores on the Personal Growth subscale (which bears an inverse relationship to the misery subscales) were considered separately.

The Integration of Stressful Life Events Scale (ISLES; Holland, Currier, Coleman, & Neimeyer, 2009) assesses the degree to which a stressful life experience has been adaptively incorporated into a broader life story that may promote a sense of internal coherence and foster a secure, hopeful view of the future. The ISLES evolved from narrative therapy contributions to the process of growth and recovery following a traumatic event (Neimeyer, 2006) where it is thought that traumatized individuals are thrust into an existential journey to find or make meaning out of a trauma/loss. The ISLES features 16 Likert-style items and yields scores on two subscales, one measuring footing in the world following the stressful life event and the other gauging the comprehensibility of the stressor, as well as a total score.

Confirmatory factor analysis substantiated the emergence of two subscales within the ISLES (Holland et al., 2009). Good internal consistency was found for both the footing in the world factor (Cronbach’s α = .93 and .94 in the general stress and bereaved samples, respectively) and the comprehensibility factor (Cronbach’s α = .80 and .85 in the general stress and bereaved samples, respectively). The total score on the ISLES likewise showed high internal consistency (Cronbach’s α = .92 and .94) in the two reported samples. Given that both subscales of the ISLES tap adaptive aspects of stress recovery, it is not surprising that they correlated with each other (r = .57 and .69, in the general stress and bereaved groups, respectively).

In terms of convergent validity for the ISLES, Holland et al. (2009) found that comprehensibility correlated strongly with a one-item measure of sense-making following stress or loss. Both footing in the world and comprehensibility correlated significantly with better mental and physical health in the general stress and bereaved samples, and with fewer symptoms of complicated grief in the bereaved sample. Thus, the ISLES appears to be a promising measure of adaptation following stress or loss and therefore an ideal scale with which to further investigate the GPI’s validity.

Finally, the Myers–Briggs Type Indicator (MBTI; Myers & McCaulley, 1985) is a measure of personality based on the psychological teachings of Carl Jung. It assesses the relative strength of four distinct psychological processes: introversion versus extraversion, sensing versus intuition, thinking versus feeling, and judging versus perception. These patterns are thought to influence how one takes in information and makes decisions (Hammer, 1993; Quenk, 1993). According to Myers and McCaulley, extroversion versus introversion describes how the individual prefers to acquire energy. The extrovert draws upon the outer world and takes in energy from external sources such as people and activities. The introvert, conversely, is energized by the inner world and prefers reflection and solitude. Sensing versus intuition focuses on how individuals acquire information. Those with a preference for sensing utilize concrete data and tangible information from the world while those with a preference for intuition utilize indirect ways of knowing, such as a “sixth sense,” in appraising a situation or person. Thinking versus feeling refers to the individual’s decision-making preferences. The thinking preference is grounded in an objective, rational method for deductive decision-making while the feeling preference operates out of a personal, subjective, and values-oriented perspective. Judging versus perceiving addresses the organizational needs of the individual. The judging preference has a high need for organization and seeks closure while the perceiver prefers open-ended flexibility and spontaneous approaches to life. According to Doka and Martin (2010), the Jungian function of thinking versus feeling has particular pertinence for their theory of grief patterns. They posit that instrumental grievers use thinking as their dominant mode while intuitive grievers use feeling as their preferred mode of functioning.

Reviewing the internal consistency of the MBTI scales from several studies, Wheeler (2001) found alpha coefficients typically in the .80s and .90s, with scores ranging from .69 to .92. His validity data on the MBTI were more mixed but factor analytic studies generally confirmed the theoretical four-scale model. Myers and McCaulley (1985) reported when individuals retake the MBTI, the percentages of agreement for the four categories are typically higher than 80%. McCrae and Costa (1989) were more pessimistic about the psychometric properties of the MBTI and suggested its constructs could be better explained in the language of their five-factor model of personality (i.e., neuroticism, extraversion, openness to experience, agreeableness, and conscientiousness).

Procedure

Interested participants from the three venues described earlier were directed to an online research website. Once participants entered the website, they read a more complete description of the research topic and answered the qualifying questions. Those individuals who met the inclusion criteria for the study were then directed to an informed consent page. Once participants accepted the conditions of the study, they completed the demographic information used to describe the study sample. Finally, they completed the four psychometric questionnaires online. The current investigators used a deidentified version of the original dissertation data which contained no protected health information and employed only a participant number for identification. All statistical analyses contained in this report are the original work of the current authors and do not duplicate any previous analyses.

Results

Reliability

The first step of the analysis consisted of deriving Cronbach’s alpha coefficients for the three GPI subscales: intuitive, instrumental, and dissonant. The authors followed the subscale assignments and item-scoring system provided by Doka and Martin (2010) to arrive at three subscale scores for each participant.

For the GPI intuitive subscale, the initial alpha coefficient of .753 indicated acceptable reliability. One item (#28 It is easy for me to put my feelings into words and discuss them with others) generated low or negative intercorrelation values with the other 10 items of the intuitive scale. Dropping this low-performing item yielded a slightly improved alpha coefficient of .769 for a 10-item version of the intuitive subscale.

Similarly for the original GPI instrumental subscale, an alpha coefficient of .716 indicated a marginally acceptable level of reliability. Again, one item (#18 I find that solving problems associated with my loss helps me) from the GPI instrumental subscale produced interitem correlation values that were low/negative. Therefore, it was dropped. The resulting 11-item version of the instrumental subscale yielded an improved alpha coefficient of .766. These calculations for both subscales indicated acceptable levels of internal consistency, that is, good reliability for the (reduced) intuitive and instrumental subscales of the GPI.

On the other hand, the 7-item dissonant subscale of the GPI showed low internal consistency (Cronbach’s α = .659) as well as low interitem correlations. These findings suggested that the dissonant items need to be interpreted individually as recommended by Doka and Martin (2010) rather than considered as a homogeneous scale.

To further investigate the GPI’s reliability, the authors explored whether individual items from the dissonant subscale might enhance the internal consistency of the intuitive and instrumental subscales. Two dissonant items (#12 and #20) showed a significant positive correlation with the (reduced) intuitive scale. Adding those two items to create a 12-item subscale, herein named Intuitive B, yielded an improved alpha coefficient of .795. The other five dissonant items showed a significant positive correlation with the (reduced) instrumental subscale. When those five items were added, the resulting 16-item subscale, herein named Instrumental B, showed a much higher alpha coefficient of .824. Thus, redistributing the seven dissonant items of the GPI to either the intuitive or instrumental subscales seemed to produce more internally cohesive iterations of these respective subscales. Also, the Intuitive B and Instrumental B subscales appeared to be independent factors as their intercorrelation was negligible (r = −.01).

These item changes from the original GPI subscales which rendered the reduced and modified versions of the intuitive and instrumental subscales are displayed in Table 1. The remaining analyses in this investigation were conducted using the Intuitive B and Instrumental B versions.

Table 1.

Grief Pattern Inventory Items With Scale Assignments.

Item	Original	Modified	Content
1	Intuitive	Intuitive B	I am more emotional than most people I know
2	Intuitive	Intuitive B	It is easy for me to cry and show my feelings to others
3	Intuitive	Intuitive B	Even though I have returned to my normal routine, I still have strong and painful feelings about my loss
4	Dissonant	Instrumental B	Even though I feel like crying, I do not cry in front of others
5	Instrumental	Instrumental B	Although I am grieving in my own way, others may think me cold and unfeeling
6	Instrumental	Instrumental B	I don’t seem to get as upset as most other people I know
7	Intuitive	Intuitive B	I feel overwhelmed by my feelings of grief
8	Intuitive	Intuitive B	I appreciate when others encourage me to share my painful feelings with them
9	Instrumental	Instrumental B	I avoid highly emotional or “touchy-freely” situations of any kind
10	Dissonant	Instrumental B	It is important to me that others view me as being in control.
11	Instrumental	Instrumental B	I have been told that I am avoiding my grief even though I do not think that I am
12	Dissonant	Intuitive B	I have been controlling my painful feelings by drinking or using other prescription or nonprescription drugs
13	Intuitive	Intuitive B	I believe that a bereavement support group is (would be) very helpful for me
14	Dissonant	Instrumental B	I worry that I am not as upset by my loss as I should be and feel guilty that I do not have more intense feelings
15	Instrumental	Instrumental B	I resent efforts to get me to show feelings that I do not have
16	Instrumental	Instrumental B	I think more about my loss than feel things about my loss
17	Intuitive	Intuitive B	I believe it is very important to be aware of, and in touch with, all of my feelings
18	Instrumental	Excluded	I find that solving problems associated with my loss helps me
19	Intuitive	Intuitive B	Although I can sometimes control my painful feelings, they usually return and overwhelm me
20	Dissonant	Intuitive B	Since my loss, I feel like I am just pretending to be strong in front of most people
21	Intuitive	Intuitive B	I find that I cannot stop my grieving by thinking of other things
22	Instrumental	Instrumental B	I have taken deliberate action to honor the memory of my loved one, even though I have not been as upset as most others who are grieving my loved one
23	Instrumental	Instrumental B	Others seemed surprised by my recovery from my loss
24	Intuitive	Intuitive B	Although I took care of things immediately after my loved one’s death, I was surprised when I eventually “crashed” and began to have intense, painful feelings.
25	Instrumental	Instrumental B	I would describe myself as more intellectual than emotional
26	Dissonant	Instrumental B	If I am upset or sad, I do not like to show it for fear that I will be seen as weak
27	Dissonant	Instrumental B	It does not usually occur to me to deal with my pain by talking about what is bothering me
28	Intuitive	Excluded	It is easy for me to put my feelings into words and discuss them with others
29	Instrumental	Instrumental B	I usually respond to the question “how are you feeling” with what I am not feeling (e.g., “not too bad”)
30	Instrumental	Instrumental B	I do not see the value of talking about feelings that I do not have in the first place

Note. Item numbers correspond to original 30-item Grief Pattern Inventory (GPI) scale proposed by Doka and Martin (2010).

Factor Structure

Next, we conducted a principal axis factor analysis of the 28 items contained in the Intuitive B and Instrumental B subscales of the GPI. This approach yielded a two-factor solution (rotated orthogonally) accounting for approximately 32% of the variance. Eigenvalues of 5.75 and 4.49 were found for Factors 1 and 2, respectively (Bartlett’s Test of Sphericity χ²= 2,040.66, df = 378, p < .001). Factor 1, the highest loading factor, corresponded essentially with the Instrumental B subscale, including negative correlations with some Intuitive B items. Factor 2 corresponded with Intuitive B as a secondary loading factor. These results corroborate that there are indeed two subscales within the GPI—instrumental and intuitive—with instrumental being the more psychometrically robust of the two scales.

Construct Validity

Having established that there are two factors within the GPI, intuitive and instrumental, and that those two (modified/expanded) subscales show good to excellent reliability, we turned to the question of construct validity. Participant scores on Intuitive B and Instrumental B were correlated with scores on the HGRC (despair, panic, blame/anger, depersonalization, disorganization, and personal growth) and the ISLES (footing in the world and comprehensibility). These intercorrelations are displayed in Table 2.

Table 2.

Correlations of Grief Variables With the Modified Scales of the Grief Pattern Inventory.

		GPI intuitive B		GPI instrumental B
Measure	Scale	Correlation coefficient	Significance	Correlation coefficient	Significance
HGRC	Despair	.644	p < .001	.220	p < .005
	Panic	.449	p < .001	.246	p < .001
	Blame	.369	p < .001	.210	p < .005
	Detachment	.484	p < .001	.316	p < .001
	Disorganization	.426	p < .001	.181	p < .01
	Overall misery	.545	p < .001	.264	p < .001
	Personal growth	.155	p < .05	.111	n.s.
ISLES	Footing	−.492	p < .001	−.204	p < .005
	Comprehensibility	−.392	p < .001	−.209	p < .005
	Total	−.478	p < .001	−.215	p < .005
MBTI	Internal-External	.084	n.s.	−.145	p < .05
	Sensing-Intuition	.079	n.s.	−.051	n.s.
	Thinking-Feeling	.320	p < .001	−.108	n.s.
	Judging-Perceiving	.061	n.s.	.021	n.s.

GPI = Grief Pattern Inventory; HGRC = Hogan Grief Reaction Checklist; ISLES = Integration of Stressful Life Events Scale; MBTI = Myers–Briggs Type Indicator; n.s. = not significant.

While both Intuitive B and Instrumental B correlated positively with the misery scales of the HGRC (i.e., higher grief pattern scores associated with higher levels of distress), individuals endorsing the Intuitive B pattern demonstrated a stronger relationship in this regard, especially on the HGRC Despair subscale as well as the total misery score from the HGRC. In other words, intuitive grievers thought to utilize a more affective way of experiencing and expressing their grief reported more negative emotions of grieving. Also, Intuitive B grievers showed a weak but statistically significant positive correlation with personal growth (p = .025). Yet, neither subscale of the GPI was strongly related to the personal growth subscale of the HGRC suggesting that neither grief pattern appeared to be decidedly more adaptive.

When considering correlations with the ISLES, similar findings emerged. Both Intuitive B and Instrumental B grievers showed significantly lower scores on footing in the world and comprehensibility, again suggesting that the more individuals endorse one of these two grieving styles, the less adaptive they seem in responding to the stress of loss. That conclusion appeared to be particularly true for those endorsing the Intuitive B pattern, but the fundamental finding remained—neither pattern was clearly more adaptive.

In order to test the limits of these grief patterns as far as their construct validity and relative adaptability, we selected participants who scored in the highest quartile on Intuitive B (n = 38; 73.7% women) and on Instrumental B (n = 37; 67.6% women) without showing an overlapping high score on the opposite scale. These relatively “pure types” comprised about 36% of the sample studied. With these more extreme cases, any differences on the HGRC and ISLES were attenuated. Intuitive B pure scorers were higher on HGRC despair (t = 3.28, df = 73, p < .01) while Instrumental B pure scorers were higher on ISLES footing in the world (t = 2.37, df = 73, p < .05). Again, neither grief pattern appeared to be convincingly more adaptive. Of interest is the fact that neither of these “pure type” groups showed a gender ratio much different from the overall sample, i.e., 69.4%.

Myers–Briggs Type Indicator

Given that Doka and Martin (2010) purport that instrumental grievers use thinking as their dominant mode while intuitive grievers use feeling, we wanted to see how the Myers–Briggs typologies related to grief styles. Taking Intuitive B and Instrumental B as continuous variables and correlating them with the results of the four subscales of the MBTI showed that intuitives scored higher as a group on Feeling (r = .32, p < .001), a finding consistent with Doka and Martin’s concepts. Interestingly, instrumentals scored slightly higher on Introversion (r = .15, p < .05) raising the possibility that instrumentals may be less inclined to openly express their grief experience or consciously “dwell” on it during interpersonal interactions.

Another exploration of the relationship between GPI scores and the MBTI involved identifying four key “functional pairs” from the MBTI (ST = sensing-thinking; SF = sensing-feeling; NT = intuitive-thinking; and NF = intuitive-feeling) as predictor variables in analysis of variance with GPI scores. Using this approach, individuals identified as ST (sensing as their perceptual mode and thinking as their decision-making mode) more often endorsed an instrumental grief style (t = .285, df = 47, p < .01). This finding seemed to suggest that instrumentals may be more cerebral overall in how they experience the grief and loss process, an interpretation clearly consonant with the Doka and Martin (2010) notion of grieving styles.

Cluster Analysis

A final effort at testing the limits of the Doka and Martin (2010) model consisted of employing participants’ scores on both the Intuitive B and Instrumental B subscales simultaneously in a cluster analysis. Four aggregate groupings or clusters emerged from our sample: low intuitive-very low instrumental (33.5%); low intuitive-average instrumental (27.8%); average intuitive-very high instrumental (14.8%); very high intuitive-average instrumental (23.9%). As a double check, we examined the gender ratio in these four clusters and found the following percentages of females: low intuitive-very low instrumental (64.3%); low intuitive-average instrumental (71.4%); average intuitive-very high instrumental (70.0%); very high intuitive-average instrumental (71.0%). Only the low intuitive-very low instrumental cluster showed a slightly lower preponderance of women.

For purposes of further analysis, we created two “superclusters” defined primarily by participants’ outcomes on the Intuitive B subscale. Supercluster 1 consisted of participants scoring low on Intuitive B (including very low or average Instrumental B scores). Supercluster 2 consisted of participants scoring average or high on Intuitive B (including very high or average Instrumental B scores).

When looking at the performance of these two superclusters on HGRC and ISLES scores, differences emerged which seemed to recapitulate the thrust and direction of previous findings. Supercluster 2 with average to high scorers on Intuitive B scored higher across the board on the HGRC distress scales, including the overall “misery” index, compared with participants in Supercluster 1 (i.e., low intuitives). The higher intuitive scorers of Supercluster 2 seemed to correlate with the conscious expression of negative grief-related thoughts and feelings. Also, Supercluster 1 with low scorers on Intuitive B produced higher scores on the two ISLES subscales of footing in the world and comprehensibility, suggesting slightly better adaptation from a meaning-making perspective among those individuals with a less pervasive intuitive grief style.

Discussion

What do the findings from this study tell us about the reliability and validity of the Grief Pattern Inventory (GPI; Doka & Martin, 2010) from a research perspective? In short, there appeared to be strong evidence for the reliability of the (modified) intuitive and instrumental subscales of the GPI and mild-to-moderate support for the construct validity of Doka and Martin’s grieving styles.

Regarding reliability, the original intuitive and instrumental subscales of the GPI showed acceptable reliability in their current form when following Doka and Martin’s (2010) original instructions. Our experimental modifications to these subscales, that is, deleting one low-performing item from each one and reassigning the dissonant items according to psychometric properties (two items redistributed to the intuitive scale and five items redistributed to the instrumental scale), produced even more reliable, robust subscales—Intuitive B and Instrumental B—based on a 28-item GPI. Both of these modified subscales showed very good internal consistency using Cronbach’s alpha coefficients. Their negligible intercorrelation indicated they were psychometrically independent of one another which verified their uniqueness and integrity as separate scales. Furthermore, factor analysis corroborated a “two-factor solution” reinforcing the presence of two independent components within the GPI.

Regarding validity of the GPI’s intuitive and instrumental subscales, the evidence was more moderate. With both grieving patterns, higher scores were associated with higher reports of grief-related distress on the HGRC. In some ways, this outcome suggested that grievers more unidimensional in their response patterns may be more prone to incurring, or at least consciously acknowledging, psychological and emotional distress following a significant loss. That inference may be particularly applicable to intuitive grievers for whom this correlation with negative grief affect was observed to be especially high.

At the same time, neither the intuitive nor the instrumental grief pattern appeared to be particularly more adaptive following loss. Intuitive grievers did score slightly higher on the HGRC personal growth subscale. At the same time, a supercluster of participants scoring low on intuitive grief registered higher scores on the ISLES factors of footing in the world and comprehensibility, both thought to align with better meaning making and better adaptation following a stressful life event, such as death of a loved one. Also, “pure” instrumental grievers scored slightly higher on the ISLES factors of footing in the world and comprehensibility. In short, there was no clear delineation of either the intuitive or instrumental pattern as the more adaptive grief style.

It was interesting that participants scoring high solely on either intuitive or instrumental grief were a clear minority in this sample. Consistent with Doka and Martin (2010), most of the study participants seemed to endorse a blend of the intuitive and instrumental grieving patterns. As a further demonstration of this phenomenon, the results of the cluster analysis showed that four basic “blends” of these two grieving patterns emerged from our sample. When participants did score high on either intuitive or instrumental grieving, the tendency was for the secondary grief pattern to be present at an average level (rather than low).

Another indication of the noncategorical nature of the intuitive and instrumental grief patterns came from an examination of the gender ratios in both the statistical “pure types” and among the four subgroups emerging from the cluster analysis. Just as Doka and Martin’s (2010) model has moved well beyond mere reductionist thinking about masculine and feminine styles of grief, our data seem to corroborate that the correlations we observed were not simply gender linked.

The results from the MBTI analysis provided mild corroboration of the constructs of intuitive and instrumental grief. Intuitive grievers were found more likely to utilize feeling as a functional pathway. Although consistent with Doka and Martin’s (2010) assertions, it remains unclear whether this finding reflects the “reporting style” of how intuitives communicate or says more about their actual experience of their grief event. Instrumental grievers were found to be more introversive and more likely to use a combination of sensing-thinking when interfacing with the world, that is, more “cerebral.” This latter finding seems to comport with Doka and Martin’s descriptions of instrumental grievers tending to “think through” the experience of grief and seeking information to cope or problem-solve in service of that style.

Of course, this study is cross-sectional and thus can offer no direct conclusions regarding causality. Nonetheless, this entire line of research is based upon an implicit notion that a person’s a priori preferential style of processing grief leads that person to selectively choose some types of postbereavement coping mechanisms over others. The results of this study can be understood and interpreted reasonably through that implicit lens.

However, it is also possible that the underlying causal frame is different. Something other than personal style, such as the raw intensity of grief affect, may play the key role in determining how persons process and attempt to cope with their grief. For example, a standard reading of our results might infer that Supercluster 2 individuals have higher misery indicators than Supercluster 1 participants because of their average-to-very-high status on both intuitive and instrumental grieving styles. An alternative interpretation could be that individuals with very intense grief affect (i.e., misery indicators) are driven by that intense grief to “pull out all the stops” and use all available expression vehicles, whether preferred or not. To ferret out this possibility, a longitudinal approach would be required so that participants’ preferred grieving style is assessed prior to an actual loss, and then their coping strategies and functioning assessed after an actual loss triggers grief emotions. Such a study would offer considerable clarity as to the causal directions that may be at work.

What can grief counselors and practitioners take away from the results of this study as far as application in clinical or other bereavement-related settings (e.g., support groups, funeral aftercare)? The intuitive and instrumental grief patterns do appear to represent internally consistent constructs that can be applied descriptively to bereaved persons and used to better identify a griever’s most reflexive response mode. However, most individuals seem to show a blend of these styles rather than an exclusive and extreme pattern.

The findings definitely support the clinical imperative to listen for and “decode” different expressions of grief following the death of a loved one. Yet, practitioners should take care not to privilege one form of grief expression over another because neither pattern showed a clear coping advantage. However, the fact that intuitive grievers tend to endorse more distress and may want to talk about their loss may result in intuitive grievers more often seeking formal clinical help following bereavement. Practitioners are cautioned to remember that this potential selection bias does not mean that everyone needs to verbally process grief-related thoughts and feelings in order to deal with loss.

Doka and Martin’s (2010) construct of dissonance in grieving did not appear to be a distinct pattern, at least as measured by the GPI. Although reflected on a content level by some of the GPI items, these items in aggregate did not constitute their own scale from a statistical perspective. Practitioners are advised to remain vigilant to the presence of dissonance for its clinical utility in helping bereft individuals who may experience internal conflicts in grief.

Certainly nothing in these results can overturn the practice axiom “one size does not fit all” in grief and bereavement. Clinicians and practitioners remain obliged to tailor their explanatory language and choice of interventions to fit a specific patient (or audience). Doka and Martin’s (2010) theory of grief patterns is only one of many conceptual “templates” that may be employed to better understand, and guide, an individual griever.

As with most empirical investigations, the current study featured certain assets as well as limitations. A weakness of previous studies investigating the GPI (Bartone & Blazina, 2016; Primeau, 2013) is the fact that the losses incurred by the participants did not involve the physical death of a significant person. Rather, those studies relied on pet loss and romantic breakups as the index event to which participants responded. The current study did investigate participants’ reactions to the death of an important person in their lives so, in that regard, provided evidence of the GPI’s reliability and validity in the primary domain upon which Doka and Martin’s (2010) concepts were founded—a clear strength.

In the authors’ view, sampling a participant cohort which represented a variety of relationships between participant and decedent also gave broader strength to the reported findings in terms of potentially wider clinical applicability. At the same time, the authors acknowledge the fact that this sample contained a higher percentage of bereaved parents (41%) than many grief studies based on a preponderance of widowed persons. Also, time since the death may be an arguable factor in terms of conceptual clarity because of the extreme range represented in the current sample, that is, 6 months to 44 years. Acknowledging these features of our sample gives both researchers and clinicians some parameters in applying the results.

The fact that only 11% of the participants had sought grief counseling of any kind in response to their loss is fairly consistent with published reports of the prevalence of complicated grief among general populations of bereaved individuals in Germany (6.7%; Kersting, Brähler, Glaesmer, & Wagner, 2011), in a meta-analysis from several industrialized nations (9.8–11.0%; Lundorff, Holmgren, Zachariae, Farver-Verstergaard, & O’Connor, 2017), and from the Yale Bereavement Study in the United States (11.9–14.2%; Maciejewski, Maercker, Boelen, & Prigerson, 2016). In that respect, the current sample seemed to be representative of grievers in general in that it did not appear to carry a higher percentage of “clinically distressed” individuals.

Finally, the current study carried two definite shortcomings that need to be acknowledged. First, the most glaring limitation was the relatively narrow demographic of the participant cohort in terms of gender (over two thirds women) and race/ethnicity (almost exclusively White), most likely an outcome of focal recruiting strategies. Clearly, a more balanced sample on these important dimensions would have been preferable and generated more confidence in applying these results to a general population of grievers. Second, given that previous studies pointed toward a connection between grief patterns and attachment style (Bartone & Blazina, 2016; Primeau, 2013), it is regrettable that a standardized measure of attachment style was not included in the instruments chosen for this investigation so that those promising leads could be further corroborated, or not, with our data. This apparent omission is perhaps best understood by recalling that the context of the original data collection was in service of investigating grief types in relationship to Myers–Briggs personality typologies rather than an investigation of the reliability and validity of the GPI per se.

In conclusion, the current study offered encouraging support for both research and clinical use of the GPI. We hope that there will be continuing refinement of the intuitive and instrumental scales themselves, as was done in this investigation, in order to maximize their psychometric properties. Most importantly, the empirical evidence linking grief patterns to distress levels, adaptation indicators, and personality style not only lends support to the original Doka and Martin (2010) conceptualization but also empowers clinicians and practitioners to more confidently apply these constructs as valid dimensions in real world situations when working with grievers who consult them for help in the wake of significant loss.

Footnotes

Authors' Note

The findings of this study were presented at the annual conference of the Association for Death Education and Counseling, Pittsburgh, PA, April 2018.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Author Biographies

Louis A. Gamino is a professor of Psychiatry & Behavioral Science at Texas A&M University College of Medicine and Baylor Scott & White Health in Temple, Texas, USA.

Kenneth W. Sewell is a vice-president for Research and Professor of Psychology at Oklahoma State University in Stillwater, Oklahoma, USA.

Lisa Prosser-Dodds is a licensed professor counselor practicing at the Lilac Center in Kansas City, Missouri, USA.

Nancy S. Hogan is a professor Emeritus of Nursing, Loyola University in Chicago, Illinois, USA.

References

Bartone

Blazina

(2016) Exploring how the human–animal bond affects men in a relational way: Attachment, loss, and gender role conflict in middle-aged and young men. In: Blazina

Kogan

L. R.

(eds) Men and their dogs: A new understanding of man’s best friend, New York, NY: Springer International Publishing, pp. 231–256.

Doka

K. J.

Martin

(2010) Grieving beyond gender: Understanding the ways men and women mourn, New York, NY: Routledge.

Gamino

L. A.

Sewell

K. W.

Easterling

L. W.

(2000) Scott & White Grief Study—Phase II: Toward an adaptive model of grief. Death Studies 24: 633–660.

Hammer

(1993) Introduction to type and careers, Palo Alto, CA: Consulting Psychologists Press.

Hogan

N. S.

Greenfield

D. A.

Schmidt

L. A.

(2001) The development and parametric testing of the Hogan Grief Reaction Checklist. Death Studies 25: 1–35.

Holland

J. M.

Currier

J. M.

Coleman

R. A.

Neimeyer

R. A.

(2010) The Integration of Stressful Life Experiences Scale (ISLES): Development and initial validation of a new measure. International Journal of Stress Management 17(4): 325–352.

Kersting

Brähler

Glaesmer

Wagner

(2011) Prevalence of complicated grief in a representative population-based sample. Journal of Affective Disorders 131: 339–343.

Lundorff

Holmgren

Zachariae

Farver-Vertergaard

O’Connor

(2017) Prevalence of prolonged grief disorder in adult bereavement: A systematic review and meta-analysis. Journal of Affective Disorders 212: 138–149.

Maciejewski

P. K.

Maercker

Boelen

P. A.

Prigerson

H. G.

(2016) “Prolonged grief disorder and “persistent complex bereavement disorder,” but not “complicated grief,” are one and the same diagnostic entity: An analysis of data from the Yale Bereavement Study. World Psychiatry 15: 266–275.

10.

Martin

T. L.

Doka

K. J.

(2000) Men don’t cry … women do: Transcending gender stereotypes of grief, Philadelphia, PA: Brunner-Mazel.

11.

Martin

T. L.

Wang

W. C.

(2006) A pilot study of the development of a tool to measure instrumental and intuitive styles of grieving. Omega: Journal of Death and Dying 53: 263–276.

12.

Myers

I. B.

McCaulley

M. H.

(1985) Manual: A guide to the development and use of the Myers–Briggs Type Indicator, Palo Alto, CA: Consulting Psychologists Press.

13.

Neimeyer

R. A.

(2004) Fostering posttraumatic growth: A narrative contribution. Psychological Inquiry 15: 53–59.

14.

Primeau, J. E. (2013). Romantic breakup and college student adjustment: Attachment dimensions, patterns of grief, and sex (Order No. 3592084). Retrieved from https://search.proquest.com/docview/1435646818?accountid=7973.

15.

Prosser-Dodds, L. (2013). The relationship between grief and personality – A quantitative study (Order No. 3560947). Retrieved from https://search.proquest.com/docview/1366090911?accountid=7973.

16.

Quenk

(1993) Beside ourselves: Our hidden personality in everyday life, Palo Alto, CA: Consulting Psychologists Press.

17.

Sanders

C. M.

Mauger

P. A.

Strong

P. N.

(1985) A manual for the Grief Experience Inventory, Charlotte, NC: Center for the Study of Separation and Loss.

18.

Wheeler, P. (2001). The Myers-Briggs Type Indicator and applications to accounting education and research. Issues in Accounting Education, 16, 125–150.