SEARCH

SEARCH BY CITATION

Keywords:

  • Social expectancies;
  • Grief;
  • Bereavement;
  • Social judgments;
  • Social norms

Abstract

  1. Top of page
  2. Abstract
  3. METHOD
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

Individuals often hold strict and erroneous expectations for how others should grieve, yet this issue has been sparsely researched. A total of 161 undergraduates rated the appropriateness of various social and emotional behaviours of a hypothetical bereaved individual as a function of the recency of the death (1 month or 1 year ago), the nature of the loss (death of a spouse or child) and the gender of the bereaved. As expected, subjects deemed it inappropriate to show positive emotions and experiences 1 month after a hypothetical death and more inappropriate to show negative reactions 1 year later and there were generally higher expectations of socially appropriate behaviour for those who lost a child rather than a spouse. Unexpectedly, there were no significant differences regarding the gender of the bereaved. This study is among the first to show experimentally the widespread expectation that grief should be experienced early and be short-lived.

There are many widespread societal expectations for how individuals should show grief. Wortman and Silver (1989) published an extremely influential paper documenting how individuals tend to hold very strict—and often incorrect—expectations for how others should grieve following the loss of a loved one. These so-called “coping myths” tend to focus on themes where emotional distress is to be expected and necessary following loss, the importance of working through one's emotions, and going through some phase of intense distress that will eventually end and allow closure and resolution about the loss. Subsequent analyses of these myths (e.g. Wortman & Boerner, 2007; Wortman & Silver, 2001) have largely corroborated both their societal presence and their general inaccuracy regarding the realities of coping with loss in an absolute sense.

Harris (2009) maintains that Western society has proscribed at least four basic societal norms that dictate how we respond and react to loss. First, there is a sense that only select individuals or social relationships are worthy of societal recognition towards the bereaved. For instance, the death of one's child is often viewed as a particularly devastating type of loss (e.g. Keesee, Currier, & Neimeyer, 2008). A related second norm considers whether the loss involves stigma such as disenfranchised grief (e.g. Doka, 1989). The third norm considers a long-standing view that grief is time-specific where it has a clear beginning and end. While the notion that one only grieves for a brief period of time immediately following a loss has been greatly challenged (e.g. Doughty, 2009), general beliefs that grief must have a sense of closure persist (e.g. Boss & Carnes, 2012). The final norm details long-standing expectations of how to show grief, especially with respect to gender where women (rather than men) are encouraged to express emotion (e.g. Martin & Doka, 2000).

Very few studies have examined, particularly with experimental means, how individuals assess their expectations for others' grief under different circumstances. The studies that have even remotely considered this issue had noticeable limitations. For instance, using a between-subjects design, Kubitz, Thornton, and Robertson (1989) presented paragraph length vignettes where the gender, general type of loss (e.g. sudden or anticipated) and the emotional reaction of the bereaved were varied; subjects were merely asked how they might like the hypothetical bereaved but perceptions of the appropriateness of the displayed grief were not assessed. Using a qualitative design with a small sample, Costa, Hall, and Stewart (2007) presented hypothetical grief vignettes in order to extract themes but lacked a clear systematic manipulation of variables.

No single study can necessarily examine the myriad of variables that might shape how an individual may rate and judge hypothetical bereaved individuals. However, this study systematically explored three of the more common, basic and important factors that may guide the decisions of perceived appropriateness of various social and emotional behaviours (e.g. attending a party, feeling happy) following bereavement: the recency of the hypothetical death (1 month vs. 1 year ago), the nature or type of the hypothetical loss (death of a spouse vs. death of a child) and the gender of the bereaved. It was hypothesised that higher ratings of inappropriateness would be given in terms of attending a party, spending time with or seeking a romantic partner, engaging in personally satisfying activities and expressing happiness in situations where the loss occurred more recently (i.e. 1 month ago) and when hypothetical targets had lost a child or were female; the reverse pattern was expected in regard to the inappropriateness of feeling sorry for oneself with respect to these three aforementioned main effects. Converse higher ratings of appropriateness were expected with respect to the previously stated five dependent measures in cases where the loss occurred longer ago (i.e. 1 year) and when the hypothetical targets had lost a spouse or were male.

METHOD

  1. Top of page
  2. Abstract
  3. METHOD
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

Participants

Shortly after receiving Institutional Review Board's (IRB) approval from my University, 161 undergraduates (55 males, 106 females) from a regional campus at a large Midwestern University were recruited to complete this study and received optional course credit. The mean and median age of respondents was 26.92 and 23, respectively (SD = 9.48, range = 18–55 years). The reported marital status (in parenthesis) of subjects was as follows: single (102), married (37), divorced (10), separated (2), widowed (2) and other (8). The number of children (in parenthesis) reported by individual subjects was as follows: none (89), one (29), two (23), three (9) and four (4).

Materials and procedure

Subjects were initially told that they were going to be completing a “Survey of Personal Judgments,” which would “involve giving personal judgments by providing ratings” and then gave their informed consent. Next, all subjects provided personal information (e.g. age, sex, marital status and number of children) and completed a study unrelated to the experiment addressed in this article. Afterwards, all subjects read: “On the next several pages, you will be asked to think about and respond to several hypothetical scenarios. Please follow the directions carefully on the next several pages. From this point on, do not look back at your earlier pages in the survey.” Over the next eight pages, subjects first read the following statement that was highlighted in a text box: “Imagine a hypothetical [man versus woman] whose [spouse versus child] died [one month versus one year] ago.” This within-subjects design asked participants eight (counterbalanced) variations regarding the gender of a hypothetical target (i.e. a man or woman), the type of hypothetical loss (i.e. the death of a spouse or a child) and the recency of the hypothetical loss (i.e. occurring either 1 month or 1 year ago). In total, there were eight distinct sentences representing each of the eight (8) possible variations. For example, the variation “Imagine a hypothetical man whose spouse died one month ago” depicts a scenario where a hypothetical male lost his spouse 1 month ago.

Afterwards, for each of the eight variations, subjects provided ratings, using a 7-point Likert scale (with endpoints representing Extremely Appropriate and Extremely Inappropriate) about a hypothetical bereaved individual. All the questions began with “To what degree would it be appropriate to…” (except for the last item that included the opening clause “In general…”) followed by the following end clauses: (a) “to attend a party or other social event?” (b) “spend time with or seek a romantic partner?” (c) “engage in personally satisfying activities,” (d) “feel sorry for oneself?” and (e) “express happiness?” For instance, the first question read: “To what degree would it be appropriate to attend a party or other social event?” The fourth measure was reverse-worded given its focus on experiencing negative rather than positive emotions (which were salient in the other four items). All five of these questions were asked for each of the respective eight (8) variations where each variation was presented to subjects on separate pages. In total, subjects provided forty (40) responses (i.e. five responses for each of the eight variations).

Upon completion of these items, subjects were debriefed, thanked for their assistance and permitted to ask questions.

RESULTS

  1. Top of page
  2. Abstract
  3. METHOD
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

In order to test the central hypotheses from this study, five separate 2 (recency of loss [1 month, 1 year]) × 2 (type of loss [child death, spouse death]) × 2 (target gender [male, female]) repeated measures analysis of variances (ANOVAs) were conducted for each of the five outcome measures. Several significant main effects for each of the dependent measures were found for the recency of loss and type of loss variables. While a few select two-way interactions were significant, none of the three-way interactions were significant. The significant results are presented in terms of each independent variable.

Recency of loss

Significant main effects were found regarding the recency of loss variable for all five outcome measures: (a) “to attend a party or other social event?” F(1, 160) = 204.74, MSE = 3.85, p < .001, (b) “spend time with or seek a romantic partner?” F(1, 160) = 215.84, MSE = 3.78, p < .001, (c) “engage in personally satisfying activities?” F(1, 160) = 127.18, MSE = 2.56, p < .001, (d) “feel sorry for oneself?” F(1, 160) = 85.29, MSE = 1.51, p < .001 and (e) “express happiness?” F(1, 160) = 115.67, MSE = 2.65, p < .001. Consistent with study hypotheses, subjects found it more inappropriate to attend a social function (Mmonth = 4.09, Myear = 2.53), spend time or seek a romantic partner (Mmonth = 4.67, Myear = 3.08), engage in personally satisfying activities (Mmonth = 3.39, Myear = 2.39) and to express happiness (Mmonth = 3.29, Myear = 2.31) 1 month rather than 1 year after a loss. Subjects found it more inappropriate to feel sorry for oneself 1 year (M = 3.56) rather than 1 month (M = 2.92) following a loss. Given that this outcome variable was reverse-worded, these findings are consistent with study hypotheses.

Type of loss

Significant main effects were found regarding the type of loss variable for four of the five outcome measures: (a) “to attend a party or other social event?” F(1, 160) = 19.34, MSE = 0.75, p < .001, (b) “spend time with or seek a romantic partner?” F(1, 160) = 101.73, MSE = 2.92, p < .001, (c) “feel sorry for oneself?” F(1, 160) = 8.07, MSE = 0.70, p < .01 and (d) “express happiness?” F(1, 160) = 9.62, MSE = 0.49, p < .01. Three of the four results offered clear support of the hypothesis that subjects would generally hold higher social expectations regarding appropriate grief reactions for those who lost a child rather than a spouse. Specifically, subjects found it more inappropriate to attend a social function (Mchild = 3.42, Mspouse = 3.20) and to express happiness (Mchild = 2.86, Mspouse = 2.74) when the hypothetical loss involved the death of a child rather than a spouse. In addition, consistent with study hypotheses, subjects also found it more inappropriate to feel sorry for oneself when one lost a spouse (M = 3.30) rather than a child (M = 3.17). There was also a significant difference in terms of whether subjects viewed it as appropriate to spend time with or seek a romantic partner such that subjects tended to endorse higher ratings of inappropriate social behaviour for those who lost a spouse (M = 4.36) rather than a child (M = 3.39). Although this finding appears inconsistent with study hypotheses, the next section considers additional interpretations of this result.

There were a couple of significant interactions between loss type and the recency of the loss. The first of these interactions indicated a sharper decline in perceived inappropriateness of seeking a romantic partner with the passage of time for those who lost a spouse rather than a child, F(1, 160) = 27.61, MSE = 1.53, p < .001. The second interaction suggested that there was a slightly sharper rise in perceived inappropriateness for those who lost a spouse with the passage of time in comparison to those who lost a child with respect to feeling sorry for oneself, F(1, 160) = 8.31, MSE = 0.45, p < .01. While neither of these interactions were predicted, partial eta squared values suggested a larger effect for the former (inline image  = 0.15) rather than the latter (inline image  = 0.05) interaction.

Target gender

Unexpectedly, there were no significant main effects for target gender for any of the outcome measures. However, there were a few select significant interactions involving gender (three of which involved recency) for three of the outcome measures. There were significant interactions between gender and recency suggesting that: it was slightly more inappropriate for men rather than women to show social engagement 1 year after a loss than at 1 month, F(1, 160) = 4.34, MSE = 0.71, p < .05; it was more inappropriate to show interest in a romantic partner for women earlier after a loss than later F(1, 160) = 6.05, MSE = 0.72, p < .05; it was more inappropriate for women to show happiness soon after a loss, whereas it was much more inappropriate for men to do so 1 year later, F(1, 160) = 6.99, MSE = 0.51, p < .01. However, like the interactions detailed earlier, none of these interactions were predicted and all of these interactions suggested fairly small effects as evidenced by their partial eta squared values (where values ranged from 0.03 to 0.04).

DISCUSSION

  1. Top of page
  2. Abstract
  3. METHOD
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

This experimental study examined the expectations that individuals have with respect to hypothetical grief and bereavement in terms of the following qualities: (a) the length of time since the loss (i.e. 1 month vs. 1 year), (b) the nature of the loss (i.e. the loss of one's child or one's spouse) and (c) the sex of the bereaved. This experiment is particularly noteworthy, given that it is among the first investigations to offer clear evidence for the widely held societal assumption that grief should primarily be experienced soon after a loss and should decrease with the passage of time.

With respect to the child–spouse loss dimension, as expected, there were generally more stringent expectations of appropriateness for those who lost a child rather than a spouse: Specifically, for those who lost a child, it was judged more inappropriate to go to a party and to feel happiness and more appropriate to feel sorry for oneself. However, subjects deemed it more inappropriate for those who lost a spouse (rather than a child) to show interest in a romantic partner. The interaction between loss type and recency of loss may provide some insight about this unexpected result such that it may be especially inappropriate to pursue spending time with a romantic partner very soon after the death of one's spouse (as doing so could be construed as particularly callous). Although any study can contain spurious effects, they are much less likely when examining a priori scientific thinking rather than post-hoc analysis (Anderson, Burnham, Gould, & Cherry, 2001). Given that this sole finding was unexpected, some caution only in regard to this result is appropriate.

While the non-significant main effects regarding the gender of the hypothetical bereaved targets were somewhat surprising, the relative lack of context about the nature of the hypothetical losses (including expressed emotions) may offer perspective about these non-significant results. Although the significant interactions involving gender may not be particularly insightful, they may suggest that subjects require additional context when making evaluations of hypothetical bereavement beyond the basis of gender alone; for instance, an emotionally stoic hypothetical female (rather than a hypothetical male) may be rated more harshly (Versalle & McDowell, 2005).

This aforementioned point highlights a limitation of this study in that it did not attempt to either portray or manipulate different categories or more nuanced situational contexts of grief even though there are several possible trajectories of bereavement, including common and chronic grief (Bonanno, Wortman, & Nesse, 2004). Future investigations should not only seek to investigate such situational contexts but also consider how well, on the basis of a hypothetical vignette, such a scenario might describe a given gender, type of loss or recency of loss. Another limitation of this study was that it utilised an exclusively Western sample where the cultural background of the participants was not systematically controlled. It is also noteworthy that subjects rarely gave the highest possible ratings of perceived inappropriateness on the 7-point Likert scale. While it is possible that subjects may truly strive to be less judgmental towards the bereaved, there is a possibility that social desirability pressures may make one less predisposed to chastise any bereaved individual (whether real or hypothetical).

Even though the literature on how individuals grieve continues to grow, there is a surprising dearth of corresponding literature detailing how general populations evaluate hypothetical (or real) bereavement and formulate various social expectancies about grief. Given the marginal literature on expected grieving styles, even though the dependent measures were simple single-item scales, this study offers a potential paradigm for future investigations to test for these different expectancies. This study also highlights both the need for greater general awareness and education about the nature of grief and its social developmental trajectory to the larger population. In doing so, select widespread misperceptions about grief may be corrected.

REFERENCES

  1. Top of page
  2. Abstract
  3. METHOD
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES