Journal of Nursing Education

Development and Study of an Instrument to Measure Role Strain

Cecile A Lengacher, PhD, RN

Abstract

ABSTRACT

The purpose of the Lengacher Role Strain Inventory (LRSI) is to assess the characteristic of role strain in female nursing students who have multiple roles (being in school, having a family, and being employed). The LRSI consists of a 100-item Likert-type scale instrument. It was administered to 327 students (a pilot and five groups) in three different colleges. Reliability and internal consistency were determined with alpha coefficients. Coefficients of .93, .94, and .95 were obtained for female nursing students and .95 and .87 were obtained for the nonnursing community college student groups. Evidence of construct validity of the LRSI may be seen in comparing the mean scores of female nursing students and male and female nonnursing students using a contrasted-groups approach. Tests of significance between the nursing student groups and the nonnursing students resulted in significant differences of p<.05, p<.01, p<.0001, and p<.0000.

Abstract

ABSTRACT

The purpose of the Lengacher Role Strain Inventory (LRSI) is to assess the characteristic of role strain in female nursing students who have multiple roles (being in school, having a family, and being employed). The LRSI consists of a 100-item Likert-type scale instrument. It was administered to 327 students (a pilot and five groups) in three different colleges. Reliability and internal consistency were determined with alpha coefficients. Coefficients of .93, .94, and .95 were obtained for female nursing students and .95 and .87 were obtained for the nonnursing community college student groups. Evidence of construct validity of the LRSI may be seen in comparing the mean scores of female nursing students and male and female nonnursing students using a contrasted-groups approach. Tests of significance between the nursing student groups and the nonnursing students resulted in significant differences of p<.05, p<.01, p<.0001, and p<.0000.

Introduction

An area of concern today for nursing students returning to school is the increasing demands of many roles expected of the student. The average female nursing student is faced with many role expectations in addition to the academic pressures of a nursing program. The availability of a reliable and valid role strain inventory would allow researchers to examine the characteristics of role strain in relationship to other variables. This characteristic is very important to women in pursuit of higher education in nursing.

According to the literature, role conflict, role stress, and role strain of women have been documented to have negative effects and positive effects. Hall (1972) identified the home role as the major source of strain and conflict for college women. Hall and Gordon (1973), Coiner (1978), and Hall (1975) described that the added pressures of working contribute to the stress of women. Feldman (1973) cited that marriage has a severe effect upon the role of women. For women returning to school, Berkove (1979) found that the support from one's husband is significantly related to less stress in women returning to school. Gray (1983) studied role conflicts and coping strategies of 232 professional married women (doctors, lawyers, professors) and found 77% of women experienced strains between home and career; however, the rewards of a career and family are well worth the effort. Van Meter (1976) identified that time pressures, workload pressures, and feeling tired are contributions to role strain in college women. Thoits (1983) identified that women who have increased roles exhibit greater distress levels than males. In contrast, Baruch and Barnett (1986) identified that the quality of the role is more important than the number of roles.

In contrast to negative effects, Van Meter (1976) found that role conflict due to school can be compensated or overcome by the feeling of "adequacy" that accompanied being in school. Campaniello (1988), in a study of role conflict and well-being of multiple-role women, found that multiple roles did not increase perceived role conflict. Nurses with more roles experienced greater well-being than nurses with fewer roles. Verbrugge (1983) found out that most people with multiple roles tend to have the best health.

In summary, 10 content areas related to role strain emerged from the review of the literature: homemaker, support of husband or significant other, school, time pressures, economic pressures, personal health, children, career, community activities, and recreational activities. The majority of studies reviewed focused on the deleterious effects of role stress, role strain, and role conflict due to increased roles of women. However, the positive effects of role strain are becoming of interest. No consistent measurement for the characteristic of role strain was found in any of the reviewed studies.

The Lengacher Role Strain Inventory (LRSI) was first tested in a descriptive study of the relationship of personality, stage of career development, and marital status to role strain (Lengacher, 1983). Results of a stepwise regression analysis showed that role strain was significantly related to all variables in the predictive model: personality, stage of career development, and marital status. Statistical validation of the model implied construct validity to the instrument. In a study of variables related to success on the NCLEX-RN, Lengacher and Keller (1990) reported that role strain (using the LRSI) was not significantly related to success on the examination, and not related to admission and exit variables identified in the program. This article describes the LRSI, an instrument that measures the dimension of role strain and the testing of the instrument that established reliability, content, and construct validity.

Theoretical Framework

The theoretical framework for this inventory is based upon Goode's (1960, 1973) theory of role strain. Goode originated his theory from Parsons and Shils (1962), Linton (1945), and Merton (1957). Merton asserts that the potential for disturbance of the "role set" is normal, due to multiple role obligations. Similarly, a major assumption supported by Goode is that multiple role obligations can be a source of strain (role strain) and that this is a condition in social life that all people face. Goode defined role strain as the felt difficulty in meeting role obligations and refers to the subjective nature of role strain. He believed that an individual faces many role demands and cannot meet all of them. Concurrent with his belief that role strain is a "normal" experience that everyone encounters, in general, he believed that an individual's multiple role relationships are over-demanding.

There have been several terms used to describe role strain: anxiety, tension, frustration, and stress. From Goode's (1960) perspective, role strain is the felt difficulty in meeting role demands, and it is normal for most people to have role strain. Role strain in this inventory refers to a subjective experience that can be described as a tension, a driving force, anxiety, and/or frustration that a woman may experience due to multiple demands she puts upon herself or demands put upon her by others. This tension or driving force can be viewed positively or negatively. As the concept of role strain is systematically studied, the definition of role strain can be refined to include new knowledge of situations related to role strain.

Methodology

Instrument design

The construction of the LRSI began with the purpose of developing an inventory that would measure role strain of the women who had multiple roles - student, mother and/or wife, and employee. The development and validation of the LRSI involved reviewing the literature to determine item content, assessing the content validity of the instrument, writing the items, piloting the instrument, attaining item analysis and reliabilities, and revising the piloted instrument. The instrument was developed considering the American Psychiatric Association's ( 1974) standards for test development, which identifies that procedures be completed for validity and reliability determination to support claims made about the test. The investigator searched the universe of content from which items could be written. A content grid, developed from a literature review, identified 10 content areas related to role strain: homemaker role, support of husband or significant other, school, time pressures, economic pressures, personal health, children, career, community, and recreation activities.

The content derived from a review of the literature was evaluated by a panel of six experts in nursing and test construction. All experts were doctorally prepared nursing faculty who had experience in teaching registered nurses. They were asked to assess the relevancy of the content to role strain by judging the content derived from the literature and by indicating the proportion of content that was related to role strain in each content area. They were asked to add categories that were not covered in the 10 content areas identified by the review of the literature.

After evaluating the content areas presented, three judges recommended that items be included related to collégial acceptance. Specific items were written and integrated into the category of career. One judge suggested combining personal health and recreational activities. A third judge felt that difficulty in meeting all role obligations would be a major factor in the amount of role strain experienced. This dimension was emphasized in the items under the content area of time. Based on their judgment, 10 categories and proportions of content related to role strain were specified. Based on expert review of the items, the following proportions were derived: time pressures, 18%; husband/significant other support, 16%; school, 16%; children, 14%; career, 12%; homemaker, 8%; economic pressures, 6%; personal health and community activities, 8%; and recreational activities, 2%. After the content areas related to role strain were identified, 100 items were written in accordance with the proportional number in each content area and were based upon item-writing criteria identified by Edwards (1957) (Figure).

After the items were written, the experts were asked to review the items for clarity and appropriateness to the category. The instrument was designed using a summated rating scale. A five-point Likert-type scale of from (1) strongly disagree to (5) strongly agree, was used for the response categories and allowed for attitude intensity. The items were coded so that the higher the response number, the higher the role strain experienced by the person. Fifty items were designed as positive items related to role strain; 50 items were designed as negative items related to role strain. After the items were developed, they were randomly distributed using a table of random numbers. The possible range of points could vary from a zero to 500; a zero indicating absolutely no role strain, and 500 indicating complete role strain choice on all the items.

Sample

The sample was composed of 327 nursing students (five normative groups plus a pilot group) in three colleges. Demographic data comparisons are identified in Table 1 for all five groups. The instrument was piloted on 41 female RN students attending a weekend baccalaureate (BSN) program in a private college. After the instrument was piloted, the revised instrument was administered to the first group of 86 female RN students attending a BSN-completion program different from the college used in the pilot group. The second and third groups consisted of 42 and 64 female nursing students in an associate degree (ADN) program in a community college.

The fourth and fifth groups were nonnursing community college students enrolled in the same community college as the ADN students. These students were identified to normally have less role strain than the nursing students at the community college level. Group 4 was composed of 68 female nonnursing community college students and Group 5 was composed of 26 male nonnursing community college students.

Procedure

All students were recruited on a voluntary basis in classrooms and the instrument (LRSI) was administered using human subjects guidelines. Informed consent was obtained from all participants.

Stressful times in each program had been previously determined through student evaluations. For the pilot group, BSN ÜV=41), the most stressful time was assessed to be at the end of the first semester prior to taking challenge examinations for lower division nursing credits. Similarly, for the BSN-completion students, Group 1 (N= 86), the most stressful time was assessed to be at the end of their first semester, prior to taking challenge examinations for lower division course work.

For the ADN students, Groups 2 OV= 42) and 3 (N=64), the most stressful time was assessed to be in the middle of their second semester of the first year of the program - the first semester of heavy medical/surgical clinical and theory content. The data on the BSN and RN-completion students were gathered prior to the data on the ADN students.

The community college nonnursing students, Groups 4 (N= 68) and 5 (N= 26), were used as the contrasted groups to determine construct validity. These students were administered the inventory during their general education classes in the summer session of their program. The data for the contrasted groups were gathered two years after data were attained on the ADN students at the same community college. The rationale for collecting data on the nonnursing community college students is that if the LRSI was a valid role strain inventory for women, then a difference should exist in a contrasted group, the male students.

FIGURECriteria for Writing Items

FIGURE

Criteria for Writing Items

Results

Cronbach's alpha was utilized to establish the reliability of the instrument. Content validity was first determined and then construct validity was determined using £ tests for independent samples for the psychometric evaluation with contrasted groups (Waltz, Strickland, & Lenz, 1991).

Reliability and internal consistency

The reliability of the piloted instrument was determined using coefficient alpha and the instrument was revised, based upon the analysis of the items in the piloted sample. Of the 100 items, 13 items had low correlations with the total scores (less than .20). Items with low correlations do not have variance related to the common factor among the items. These items decrease the test reliability (Nunnally, 1978). The 13 items that had a .20 correlation or below were evaluated and revised for clarity and readability.

Table

TABLE 1Demographic Data Comparisons

TABLE 1

Demographic Data Comparisons

Table

TABLE 2Means, Standard Deviations, and Alpha Coefficients for Groups of Female Nursing Students

TABLE 2

Means, Standard Deviations, and Alpha Coefficients for Groups of Female Nursing Students

Table

TABLE 3Means, Standard Deviation, and Alpha Coefficient for Contrasted Groups (Non nursing Community College Students)

TABLE 3

Means, Standard Deviation, and Alpha Coefficient for Contrasted Groups (Non nursing Community College Students)

Reliability measures used data collected on each of the three intended samples and two contrasted groups. The means, standard deviations, and alpha reliability coefficients of the pilot group and three nursing student groups are reported in Table 2, and the two nonnursing community college groups in Table 3. Means for the six groups including nonnursing students varied from a low of 246.03 to a high of 274.22. The lowest mean was reported for the nonnursing community college student group. The alpha coefficient varied from .87 to .95 in the six samples. The nonnursing male group had the lowest reliability coefficient (.87X

After alpha coefficients were derived for each group, items were individually analyzed to identify items with negative or low magnitude item-to-scale correlations. Nunnally (1978) identified that deletion of these items could result in higher internal consistencies. However, the number of items having low and high correlations with the total test scores varied and were not consistent across groups. Each group was assessed for items that had low correlations (negative to .20) with the total test score: Group 2 (N= 42) had the highest number of items - 23; Group 3 (N= 64) had 19 items with low correlations; and Group 1(W=86) had nine items. For Group 4(N= 68), the number of items that had low correlations was 14; for Group 5 (JV =26), the number of items was 38. It is evident that the nonnursing male group of students reported more items that did not have high item-scale correlations.

The following are examples of items having a low correlation with the total test score: "Returning to school is not vital to my future career, and the strain is not worthwhile," and 'Going back to school will not give me future security."

Each group was also assessed for items having high correlations with the total test score (>.60): Group 2 (JV =42) had eight items with a high correlations, Group 3 (N= 64) had five, and Group 1 (N= 86) had four items. For the nonnursing community college groups, Group 4 (N= 68) had six items with high correlations and Group 5 (N= 26) had only four items.

Examples of items having high correlation with role strain are: "My relationship with my family/children has deteriorated since I have returned to school," "I find it exhausting to continue with all my household obligations in addition to studying and working," and "My family/ friends do not give me any emotional support." In summary, the internal consistency of the LRSI was quite high as measured by coefficient alpha and resulted in high reliability coefficients.

Construct validity

Construct validity was determined using the contrastedgroups approach. Factor analysis was not considered because of the small sample size currently available (Nunnally, 1978). Waltz et al. (1991) identified that construct validity can be determined using the contrastedgroups approach or what Polit and Hungler (1991) call the known-groups technique. Waltz et al. reported that if the investigator can identify two groups of individuals who could be extremely high and extremely on the characteristic, and if the instrument is sensitive to the individual differences, statistical significance should be reported through the t test or analysis of variance. The investigator identified groups of individuals who would be expected to differ in the characteristic measured (female nursing students and nonnursing students, males and females). Waltz et al. indicated that the investigator may claim evidence for construct validity if the instrument is sensitive to the trait being measured and if differences occur between groups. Similarly, Polit and Hungler identify that there would be an anticipation of some group difference in the scores attained.

Evidence of construct vali dity of the LRSI can be seen in comparison of mean scores of each group. Differences can be observed in Table 2 where the mean scores of the female nursing student groups were higher, ranging from 253.52 to 277.61, compared to Groups 4 and 5 in Table 3 where means were lower, ranging from 246.03 to 247.15. This suggests that the nursing student groups experienced higher role strain than nonnursing college students, including males. In addition, the standard deviations were lower for Group 5, all males (Table 3).

Evidence for construct validity of the LRSI was revealed (Table 4) when there was an examination of the mean score differences between the groups that included nursing and nonnursing community college students: t tests for independent samples were completed between Groups 3 and 4, Groups 3 and 5, Groups 2 and 4, and Groups 2 and 5 to identify what significant differences might occur between groups. These resulted in significant differences between groups ofp<.05,p<.01,p<.0001,andp<.0000.

Upon analysis of the item correlations with the total test score for nonnursing students, a substantially greater number of items were reported with low correlation to the total test score (negative to .20) in Groups 4 and 5 as compared to the three groups of female nursing students. Items with low correlation do not have variance related to the common factor, therefore assuming that role strain is not being measured.

Discussion

The LRSI instrument, which measures the characteristic of role strain, has been developed using the APA's (1974) standards for test development that identify that reliability and validity determination are essential for all tests. Content used in designing the instrument was obtained from a review of the literature related to multiple roles of women: women who work, have a family, and go to school. Content validity was determined first from the literature review, then validated by a panel of experts.

Table

TABLE 4Construct Validity (Contrasted Groups) Test of Significance

TABLE 4

Construct Validity (Contrasted Groups) Test of Significance

Overall, the LRSI performed well psychometrically for all groups tested. As expected, it did not perform as well for the male nonnursing students for which the inventory was not designed. Evidence of reliability was demonstrated through the use of coefficient alpha. Differences in acceptability of alpha coefficients vary. Nunnally (1978) identifies that .60 is acceptable for survey research; however, for decisions about individuals, Polit and Hungler (1991) identify that a .90 or better is recommended. After data were gathered on the BSN and ADN nursing students, the investigator identified a group of nonnursing community college students who were identified to be extremely low in the characteristic measured. These were female and male community college students enrolled one year later in the same community college as the ADN nursing students. Data were gathered on these nonnursing college students for evidence of construct validity. Variances in reliability estimates differ slightly between BSN and ADN nursing students from .93 to .95. The reliability index held up with the nonnursing female community college student at a .95. However, as expected, the nonnursing male community college student's reliability coefficient was lower, .87.

Construct validity is evidenced by comparison of nonnursing groups with nursing groups by utilization of i tests for independent samples that resulted in significant differences between groups. Variances in mean score differences between groups were reflected in statistical differences between groups using t tests. Low mean scores were apparent in nonnursing community college students, 246.03 to 247.15, as contrasted with the nursing students' mean score of 253.52 to 277.61. The results of this study suggest that the LRSI is a useful research tool, and that additional data should be obtained with different groups.

A reliable and valid instrument that assesses role strain is essential to contribute to research related to multiple roles of women. This instrument will enable educators to test models for application and interventions in the academic environment.

Limitations

The number of items in the inventory ( 100) should be shortened while maintaining a high reliability estimate. Further validity studies could be completed; however, a factor analysis of all 100 items was not possible because of sample size. Therefore, the contrasted-groups approach was used to assess construct validity. Nunnally (1978) identified a criteria of 10 subjects per scale item for factor analysis. The sample size does not meet the recommended subjects for the 100-itera inventory. Factor analysis could be implemented in the future if the sample size is larger or if the inventory length is decreased. Data on the BSN students were gathered prior to data on the ADN students. Data on the nonnursing community college students were collected after the data were gathered on all groups of nursing students. The nonnursing community college students attended the same community college as the ADN students but were enrolled in the college one year later. For future research on the instrument, refinement of items related to the length of the inventory should be a major consideration. Continued collection of data will provide a sufficient sample for factor analysis to further increase the tool's usefulness and validity.

Utilization of the LRSI

The LSRI can be used in identifying role strain of women from various academic settings. Based on the findings in this study, practical use has been documented for BSN-completion students and ADN students.

This inventory meets a definite need for assessment of women's role strain, particularly related to women who are returning to school, who are employed, and who have a family. There are no valid and reliable inventories that assess this dimension. The original intent of the instrument was to assist educators and students to assess their role strain in relationship to the multiple roles assumed by women today in our society. Assessment of role strain in students enables educators to assist students in their transition toward meeting their career goals. Action strategies and interventions can be employed so that students can meet their career goals and prevent problems related to role strain. If students are assessed to have high role strain, coping strategies can be initiated to prevent emotional and physical problems. Role strain can be a basis for future study of the many women who work, have a family, and are in school.

The LRSI is presented so that readers can utilize it in working with students to improve teaching or for research purposes. Sharing data will further refine the inventory. The future refinement also allows for realistic clarification of the concept of role strain. The major limitations of the inventory to date have been the type of groups studied: BSN-completion, ADN, and nonnursing community college students. A goal for future study is to expand this to other groups. The concept of role strain for women is a vital concern not only in nursing but in the study of women in society.

References

  • American Psychological Association. (1974). Standards for educational and psychological tests. Washington, DC: Author.
  • Baruch, G., & Barnett, R. (1986). Role quality, multiple role involvement, and psychological well-being of mid-life women. Journal of Personality and Social Psychology, 51, 528-585.
  • Berkove, G.E (1979). Perceptions of husband support by returning women students. Family Coordinator, 28, 451-457.
  • Campaniello, J. (1988). When a professional nurse returns to school: A study of role conflict and well-being of multiple-role women. Journal of Professional Nursing, 4, 136-140.
  • Coiner, M.R. ( 1978). Employment and mother's emotional states: A psychological study of women reentering the work force. (Doctoral dissertation, \ale University, 1978). (University Microfilms No. 790-15, 806)
  • Edwards, A. (1957). Techniques of attitude scale construction. New York: Appleton-Century-Crofts.
  • Feldman, S. (1973). Impediment or stimulant? Marital statue and graduate education. In J. Humber (Ed.), Changing women in a changing society (pp. 220-232). Chicago: University of Chicago Press.
  • Gray, J. (1983). The married professional woman: An examination of her role conflicts and coping strategies. Psychology of Women Quarterly, 7, 235-243.
  • Goode, W.J. (1960). A theory of role strain. American Sociological Review 25, 483-496.
  • Goode, W. J. (1973). Exploration in social theory. New York: Oxford University Press.
  • Hall, D.T. (1972). A model of coping with role conflict: The role behavior of college educated women. Administrative Science Quarterly, 17, 471-489.
  • Hall, D.T. (1975). Pressures from work, self, and home in the life stages of married women. Journal of Vocational Behavior, 6, 121-132.
  • Hall, D.T., & Gordon, RE. (1973). Career choices of married women: Effects on conflict, role behavior, and satisfaction. Journal of Applied Psychology, 58, 42-48.
  • Lengacher, C.A. (1983). Role strain as a function of personality and career characteristics. (Doctoral dissertation, Marquette University, 1983). Dissertation Abstracts International, 44, 1019A-1120A.
  • Lengacher, C.A., & Keller R. (199OX Academic predictors of success on the NCLEX-RN examination for associate degree students. Journal of Nursing Staff Development, 29, 163-168.
  • Linton, R. (1945). The cultural background of personality. New York: Appleton-Century-Crofts.
  • Merton, R.K, (1957). The rote-set: Problems in sociological theory. British Journal of Sociology, 8, 106-120.
  • Normally, J.C. (1978). Psychometric theory (2nd ed.). New York: McGraw-Hill.
  • Parson, T., & ShUs, E.A. (Eds.). ( 1962X Toward a general theory of action. Cambridge, MA: Harvard University Press.
  • Polit, D-, & Hungler, B. (1991). Nursing research principles and methods (4th ed.). Philadelphia: J.B. Lippincott.
  • Snoek, J.D. (1966). Role strain in diversified role sets. American Journal of Sociology, 71, 363-372.
  • Thoits, P.A. (1983). Multiple identities and psychological wellbeing: A reformation and test of the social isolation hypotheses. American Sociological Research, 48, 174-187.
  • Van Meter, M.J. (1976). Role strain among married college women. (Doctoral dissertation, Michigan State University). (University Microfilms No. 79-18, 683.
  • Verbrugge, L. (1983). Multiple roles and physical health of women and men. Journal of Health and Social Behavior, 24(3), 16-30.
  • Waltz, C-, Strickland, O., & Lenz, E. (1991). Measurement in nursing research (2nd ed). Philadelphia: F.A. Devis.

TABLE 1

Demographic Data Comparisons

TABLE 2

Means, Standard Deviations, and Alpha Coefficients for Groups of Female Nursing Students

TABLE 3

Means, Standard Deviation, and Alpha Coefficient for Contrasted Groups (Non nursing Community College Students)

TABLE 4

Construct Validity (Contrasted Groups) Test of Significance

10.3928/0148-4834-19930201-08

Sign up to receive

Journal E-contents