Inspired by the American Health and Retirement Survey (HRS) of adults aged 50 and over, which began in 1992, several other countries have since mounted similar efforts in order to better understand the aging of their populations and in order to better inform public policy. The Mexican Health and Aging Study (MHAS), the English Longitudinal Study of Ageing (ELSA) and the Korean Longitudinal Study of Aging (KLOSA) are notable examples. Europe joined the HRS/ELSA legacy in 2004 when it began a cross-national survey of adults aged 50 and over, and their spouses regardless of age, in 11 countries, in the Survey of Health, Aging and Retirement in Europe (SHARE) (Boersch-Supan et al. 2005). Israel joined the SHARE project in 2005, becoming the first country in the Middle East to initiate systematic study of its aging population in the HRS tradition. SHARE-Israel is supported by the NIA, the German-Israeli Fund for Development and Research and the Israeli National Insurance Institute. As part of this special issue of the Journal of Cross-Cultural Gerontology, this article is primarily focused on the diversity among the peoples of the Middle East, although issues of family and social change, especially among the recent immigrants from the USSR, are also touched on.

This paper provides an introduction to the SHARE-Israel venture. It focuses on two main themes: (1) description of the unique requirements for construction and implementation of the survey in this particular setting, and (2) presentation of preliminary results from an early release from the first wave of the survey, comparing populations within Israel and across Europe on selected parameters in the areas of health, employment and income. These topics were chosen because they are all strongly related to well-being in late life. The paper begins with a description of the setting in which the SHARE-Israel survey was implemented.

The Setting

Although Israel’s population is relatively young compared to other developed countries, its rate of aging has been faster. While the general population increased 3.7 times since 1955, the population 65 and over increased some 7.7 times. The rate of increase of those aged 75 and over was even greater—11.5. At the end of 2006, Israel had a population of 7.054 million, 919,200 of whom were age 60 or over (13%), and 697,600 (9.9%), age 65 or over. It is estimated that the proportion of elderly people in the population will remain stable up to 2010, but will reach 12% of the total population by 2020 (numbering 1,025,800 persons) (Brodsky et al. 2007; CBS 2007).

The story of aging in Israel is also a story of immigration (Brodsky and Litwin 2005; Habib and Tamir 1994). Seventy eight percent of Israelis age 50 and over were born in other countries. Moreover, many of them immigrated relatively late in life. Immigration in old age has been noted to constitute a double jeopardy, presenting unique challenges regarding language acquisition, employment and income security, housing and a range of other dilemmas in one’s twilight years (Litwin 1995). Of special note in this regard is the recent large-scale immigration, from the former Soviet Union (FSU), which began at the end of 1989. From 1989 to 2006, about a million immigrants arrived from the FSU. The immigrant population from the Former Soviet Union (since 1989) currently accounts for some 20% of persons age 50 and over.

Arab-Israelis, who include Moslems, Christians, and Druze, currently comprise about 19 percent of the Israeli population (Litwin 2005). However, Arab-Israelis constitute just 10% of persons aged 50 and over. Moreover, persons age 65 and over comprise only about 3% of the Arab population in Israel. Arabs in Israel constitute a minority population that is characterized by low socioeconomic status, early exit from the labor force among men and low rates of employment among women, traditional familial culture and familial values regarding care of the aged (Al Haj 1995; Azaiza and Brodsky 1996; Lowenstein and Katz 2000). In contrast, the Jewish majority maintains more Western values and orientations, on average, in regard to elder care. In order to reflect these varied population dynamics, the SHARE-Israel survey established parallel mechanisms to address the three major population groupings: (1) veteranFootnote 1 Jewish-Israelis, (2) Arab-Israelis and (3) recent immigrants (since 1989) from the FSU.

The Structure of SHARE-Israel

The SHARE-Israel survey is based upon the generic questionnaire that was applied in all the SHARE countries. The questionnaire is administered through a computer assisted personal interview (CAPI) that utilizes Blaise as the operating language, and through a drop-off supplement on paper. We quickly learned that the adaption and translation of the CAPI mechanism into Hebrew, Arabic and Russian for the Israeli survey presented a major challenge. In contrast to the European languages of SHARE, Hebrew and Arabic have a right to left direction. Moreover, each of the three languages in the Israeli version required the incorporation of unique fonts. The Arabic fonts proved to be the relatively most complex. These technical difficulties required special efforts, but were eventually overcome through programming. The CAPI mechanism that was developed for Israel was successfully tested in pilot interviews and subsequently adopted for implementation.

SHARE is an interdisciplinary undertaking. Areas addressed in the survey included demographics, physical health, behavioral risks, cognitive function, mental health, health care, employment and pensions, grip strength, walking speed, children, social support, financial transfers, housing, household income, consumption, assets, activities and expectations (Boersch-Supan and Juerges 2005). The Israeli survey instrument addressed two additional domains not included in the generic SHARE questionnaire: (1) the measurement of life long trauma, including exposure to the Holocaust, and (2) examination of reactions to pension reform, that is, the delay of eligibility for retirement benefits in Israel to age 67 for men, and to age 64 for women.

Since the SHARE-Israel questionnaire was administered in three different languages, as noted earlier, it was necessary to guarantee that each of the languages measured the several different areas delineated above in comparable ways. That is, we had to capture the meaning of the questions across all three languages, and within their cultural context. Feedback from native speakers and the results of pilot tests helped reformulate several questions so that they became more comparable. However, examination of the actual extent of harmonization of specific questions across the different language versions would require a full separate analysis.

The collection of data in SHARE-Israel extended over 10 months, from October, 2005 thru July, 2006. The collected data were transferred bi-weekly to the offices of CentERdata in the Netherlands, according to SHARE protocol. This included transmission of interviewer-related data. Key stroke analysis of such data allowed the ongoing monitoring of interviewer performance, early identification of sub-standard interviewers, enhancement of interviewer training and improvement of methods of respondent recruitment.

The Sample

The target of SHARE-Israel was to achieve an effective sample of 1,700 households, (about 2,500 individuals). Following HRS and SHARE practice, data were collected on all persons in the household age 50 and over and their spouses, regardless of age. Towards this end, a multi-stage design was employed combining the probability sampling of statistical areas with the creation of a sampling list within each statistical area using the telephone directory data-base. In the first stage, a sample of 150 statistical areas was drawn from the 2,300 statistical areas into which Israel is divided. The probability of inclusion was proportionate to the number of residents aged 50 and over in the statistical area. Prior to drawing, the statistical areas were stratified according to social and demographic characteristics based on nationality (Jewish and Arab areas), geographic region, type of community and socio-economic level.

In the second stage, street segments in each of the selected statistical areas were listed and matched to the national residential telephone directory file. From this list of all housing units with a phone listing in a given statistical area, a fixed number of housing units was drawn (phone coverage in Israel is approximately 95%). All units were then contacted to verify whether a person age 50 or older resided in the household. In the final stage each interviewer received a list of addresses and was instructed to interview all eligible persons.

In calculating the number of households that needed to be drawn in order to achieve the desired sample size we assumed intra-cluster correlation (ICC) of 0.02 within statistical areas (based on previous experience with the European Social Survey) and that 45% of households in Israel include at least one person 50 years and over. Taking these parameters into account and aiming for a 70% response rate, it was necessary to initially select a list of 38 addresses in each statistical area (12/[0.7 * 0.45]). Following the practice adopted in the larger SHARE project, institutional residential facilities were not included in the baseline sample of SHARE-Israel.

Preliminary Analysis

A preliminary release of data collected through the end of March, 2006, allowed initial analysis of the data in order to examine data quality and to identify possible trends and differences. The preliminary sample included 1,813 individuals (about 72% of the target sample), of which 1,694 were aged 50 or older, who resided in about 1,200 households. The current analysis focuses on the group of respondents aged 50 plus. Representation of the majority group of veteran Jewish Israelis in the preliminary sample was close to its actual proportion in the population. However, Arab-Israelis were over-represented (interviewing in this sector had been completed by end of March, 2006), and the immigrants from the FSU were under-represented. In order to rectify this deviation, the preliminary partial sample was adjusted by population group (veteran Jewish-Israeli, Arab-Israeli and immigrants from the FSU), by gender and by age group.

The data from the preliminary release were handled by the central SHARE authorities. This included initial processing, computation of scores on standard measures and treatment of missing data. In the latter case, particularly in relation to economic indicators, imputations were incorporated through the application of conditional hot deck procedures (Brugiavini et al. 2005; Christelis et al. 2005) . The current analysis looks at selected variables in the areas of health, employment and income. It must be stressed that this is a preliminary partial release of the Israeli data. As such, the findings should be considered with some degree of caution.

First reported is the area of health. Mental health was measured by the Euro-D depression scale (Prince et al. 1999). The Euro-D has been compared to the Composite International Diagnostic Interview short form which has been used by the HRS, and to similar assessment instruments (Castro-Costa et al. 2007). Depression, or Euro-D caseness, was determined in the current analysis as having received a score above the clinical cut off point of 3. Long term health problems and activity limitations were obtained by self report; respondents indicated whether or not they had either. Health behaviors examined in the current analysis included being overweight, as determined by body mass index, current smoking, and engaging in physical activity. In each case, the data were compared by population group in Israel, and in relation to the corresponding European data. The cross-country comparison serves to view the Israeli findings in relative context.

Area of employment was addressed as respondents’ work status. Respondents indicated whether they were currently employed, retired, unemployed, disabled, full time homemakers, or some other status. These figures were considered by gender. Here too, the data were compared by population group in Israel, and in relation to the European data. This comparison allows examination, for example, as to whether the early retirement trends that are prevalent in Europe exist in Israel as well, and if so, to what degree (Siegrist et al. 2007).

Finally, the current analysis considered household income. Gross total household income was computed as the sum of gross total individual income, the sum of the gross incomes of other household members and other benefits, capital assets income and rent payments received plus computed rents. These derived income measures were computed in terms of New Israel Shekels (NIS) that were subsequently converted into their Euro equivalents. Then, the Euro amounts were modified to reflect purchasing power parity (PPP). The derivation of household income included imputations on amount variables by means of information obtained from unfolding brackets. In cases in which respondents declined to provide an amount but the associated unfolding brackets provided enough information to identify an interval, conditional hot deck procedure was employed to produce imputations (Brugiavini et al. 2005; Christelis et al. 2005).

Also reported in the current analysis are selected demographic characteristics of the respondents in the preliminary sample. These include age, gender, the number of living offspring and whether or not a family dwelling is owned by the respondent. The demographic data are presented first.

Preliminary Findings

Table 1 shows the demographic data of the adjusted sample for Israel in general, and for each of the population groups separately. The mean age in the weighted sample was 63.3 (n = 1,694; range = 50–93). Analysis of variance showed that the age distribution differed by population group, with Arab-Israeli respondents being the youngest (n = 149; range = 50–93, median = 61.2) and the FSU immigrants being the oldest (n = 344; range = 50–86, median = 64.2). Veteran Jewish-Israelis reflected the overall distribution ((n = 1201; range = 50–93, median = 63.3). In terms of gender, the table shows that the FSU immigrant sub-sample had more female respondents, and that the Arab-Israeli sub-sample had fewer, but the chi square indicated that these differences were not significant.

Table 1 SHARE-Israel Preliminary Release: Selected Demographics
Full size table

The average number of surviving children was about three for the preliminary Israeli sample as a whole. Veteran Jewish-Israelis had about three surviving children, Arab-Israelis had almost seven, and FSU immigrants had between one and two. As for home ownership, the vast majority of veteran Jewish-Israelis and Arab-Israelis owned their own dwellings, 85% and 95% respectively. However, only a minority of FSU immigrants, a bit more than a third, owned a home.

The next table expands the analysis to include comparative examination of the Israeli data with corresponding data from the European SHARE countries (Table 2). The data for the ten SHARE countries in Europe were drawn from Boersch-Supan et al. 2005. Data for the preliminary Israeli sample as a whole are indicated by the letters IL. Veteran Jewish-Israelis are represented by the letters HW (Hebrew), Arab-Israelis by the letters AC (Arabic) and FSU immigrants by the letters RN (Russian). The figure reported in the first part of Table 2 is the percentage of respondents who scored above the clinical cut-off point on the Euro-D scale, by gender. The data show that the percentage of depressed women aged 50 and over in Israel (41.9%) was slightly above the corresponding percentage in the three European countries with the poorest mental health among women: France, Italy and Spain. Among the men, Israelis reported the highest percentage of respondents above the clinical cut-off for depression (27.6%). Moreover, comparison of the differences across population groups within Israel showed that immigrant women from the FSU were more depressed. Among the Israeli men, on the other hand, the differences across the population groups were not significant.

Table 2 Depression, Long-Term Health Problems and Activity Limitations among Men and Women Aged 50 Years and Above
Full size table

Table 2 also presents the data on self-rated long-term problems and activity limitations. These figures also include the average rate for all ten SHARE countries reported. First of note is that Israeli women had a ranking of long term problems that was slightly above the European average. Moreover, differences were evident among the Israeli women. The FSU immigrants reported the greatest degree of problems, and veteran Jewish-Israelis, the least. Israeli men, on the other hand, reported the greatest degree of long term problems in comparison to all SHARE countries. Differences were also evident among Israeli men from different population groups. Immigrants from the FSU had the most reported long term problems, and Arab-Israeli men, the least.

The data regarding activity limitations show some similar trends. Among both Israeli men and women, FSU immigrants were the most limited and veteran Jewish-Israelis, the least. Comparing the figures to the European average shows, in general, that Israeli women were as limited as their European counterparts. Israeli men were more limited.

Figures 1, 2, 3, 4, 5, and 6 present data on health behaviors, that is, health risk and health promotion. Figure 1 compares the scores for men in the Israeli sub-groups and the SHARE average on three indicators: being overweight, current smoking and engagement in physical activity. In regard to the latter, the figure reflects the percentage reporting relative inactivity. As may be seen, veteran Jewish Israeli men were less overweight than Arab-Israelis and FSU immigrants and less overweight than their European counterparts as well. However, veteran Jewish Israeli men reported the highest degree of current smokers compared to the other Israeli groups. On the other hand, all the male Israeli population groups reported fewer current smokers than the SHARE average for European men. In terms of physical activity, Arab-Israelis were less active than their Israeli counterparts, but all Israeli men were less active than European men, as reflected in the SHARE average.

Fig. 1
figure 1

Health behavior among Israeli men aged 50 years and above by group compared to their SHARE counterparts

Full size image
Fig. 2
figure 2

Health behavior among Israeli women aged 50 years and above by group compared to their SHARE counterparts

Full size image
Fig. 3
figure 3

Overweight/obesity among women aged 50 years and above by country

Full size image
Fig. 4
figure 4

Overweight/obesity among men aged 50 years and above by country

Full size image
Fig. 5
figure 5

Low physical activity among women aged 50 years and above by country

Full size image
Fig. 6
figure 6

Low physical activity among men aged 50 years and above by country

Full size image

The comparative data for women, as shown in Fig. 2, present a slightly different picture. Israeli women, regardless of population group, had a higher tendency for overweight than European women, on average. Arab-Israeli women were the most overweight and were the most inactive. More veteran Jewish-Israeli women and Arab-Israeli women were current smokers than European women, on average. Finally, all Israeli women were less physically active than the SHARE average in Europe.

Figures 3 and 4 show the degree of women and men who were overweight in each of the SHARE countries, including Israel. Israeli women were more overweight than women in all SHARE countries, except for Spain and Greece. In contrast, Israeli men were less overweight than their European counterparts, except for the men in Sweden, Denmark and Switzerland.

Figures 5 and 6 show the degree of women and men who were physically inactive. As may be seen, more Israeli women were inactive than women in all the comparative countries, except for Italy. Israel men, on the other hand, were the most inactive, with only Italian men approaching the Israeli rate of physical inactivity.

Work status is presented, by gender, for the Israeli sample and for the other SHARE countries, in Table 3. Several trends may be discerned. First, significant differences are apparent by population group for both women and men in Israel. Among the women, the major difference is that homemakers constituted the vast majority of the female Arab-Israeli population, aged 50 and over (89%), compared to few or none among the other groups. In contrast, almost a third of veteran Jewish women were employed and more than 40% were retired. The major difference discernable among FSU immigrant women was that almost a fifth of them were disabled, and a sixth, unemployed.

Table 3 Self Reported Labor Market Status by Gender and Country (Percentage Values)
Full size table

Table 3 also shows that the rate of employment among veteran Jewish women in Israel was among the highest in all the SHARE countries, surpassed only by the women in Sweden, Denmark and Switzerland. Also of note is the relatively high rate of Israeli women who were disabled (7.5%), a rate surpassed in Europe only by Dutch women. Finally, the percentage of women who were unemployed in Israel is higher than in any European SHARE country.

Differences in work status were also seen when comparing the respective groupings of Israeli men. Table 3 shows that only about 31% of Arab-Israeli men were employed. This is in contrast to veteran Jewish-Israeli men, 46% of whom were employed, and to FSU immigrant men (48%). On the other hand, about a fifth of Arab-Israeli men were disabled, compared to only about 8% of veteran Jewish-Israeli men and some 6% of FSU immigrant men. In addition, more than one tenth of Arab-Israeli men and FSU immigrant men were unemployed as opposed to only about 3% of veteran Jewish-Israelis. Comparing the overall Israeli average in each status to that in the SHARE countries, it appears that Israeli men had the lowest rate of retirees. Correspondingly, perhaps, Israeli men also reported the highest degree of disabled, followed by the men in the Netherlands.

The final comparison relates to gross household income, which was measured in Euro values and corrected for purchasing power parity (Table 4). Significant differences were observed among the Israeli population groups both in terms of average gross household income and median gross household income. As may be seen, the average yearly household income of veteran Jewish-Israelis was some two and a half times greater than among Arab-Israelis and FSU immigrants, before and after correction for PPP. In comparison, the median yearly household income of veteran Jewish-Israelis was about two times greater than among Arab-Israelis and FSU immigrants.

Table 4 Gross Household Income by Country (in Euros)
Full size table

In comparing Israeli gross household income to the same in the SHARE countries, a number of observations stand out. First, the average yearly household income of veteran Jewish-Israelis was about the same as the average household income for all ten SHARE countries. However, whereas PPP corrected household income in Europe reflected the unadjusted SHARE average, by design, the PPP corrected household income for veteran Jewish-Israelis was 23% higher than the unadjusted average for this same group. Table 4 also shows that the unadjusted overall Israeli average was one sixth lower than the unadjusted SHARE average, but the PPP corrected Israeli average was slightly more than the PPP corrected SHARE average, due to lower prices in Israel.

Comparing the median household incomes shows that the unadjusted medians for Israel and for SHARE are about the same. However, while the PPP corrected median for SHARE remains the same as the unadjusted one, by design, the corrected median for Israel increases by about 20%. Table 4 also shows that PPP corrected median household income among veteran Jewish-Israelis is about 60% greater than the corresponding median for the SHARE countries as a whole. However, PPP corrected median household income among Arab-Israelis is only 80% of that reported in the SHARE countries and among FSU immigrants it is less than three quarters.

The Major Trends

It should be recalled that the analyses reported in this paper are based upon a preliminary partial release of the SHARE-Israel data. Conclusions stemming from these analyses, therefore, are tentative and must be considered with some caution. Nevertheless, it is possible to identify several initial trends that will require further scrutiny.

Comparing the population groups within Israel, it appears that the majority group of veteran Jewish-Israelis aged 50 and over is in a favorable position in terms of health status, health behaviors, employment status and household income. They report the lowest degree of limitations and, among the women, the fewest long term problems. The men in this group are the least overweight. Moreover, more women in the veteran Jewish-Israeli group have a work history than in the other groups. Finally, household incomes in this population are at least twice as high as in the other groups.

Arab-Israelis aged 50 and over are at greater risk of disadvantage. A third of the men in this group are disabled or unemployed. Arab-Israeli women are the most overweight among Israeli women and the most physically inactive. Moreover, the vast majority of Arab-Israeli women are homemakers, having seven surviving children, on average. Lower household incomes in this population group may be offset, however, by the security of home ownership. Almost all Arab-Israelis aged 50 and over own their own homes.

Recent immigrants from the former Soviet Union are, perhaps, at greatest risk of disadvantage. Health-wise, they report having the greatest degree of depression, long term problems and activity limitation. They also have the fewest surviving children and low rates of home ownership. In terms of employment, about a fifth of the women in this population group are disabled and a sixth, unemployed. Finally, household income among the FSU immigrants is the lowest among adults aged 50 and over in Israel. Immigration in later-life apparently presents several obstacles to aging well.

Viewing the Israeli population of older adults vis a vis their European counterparts, several trends emerge. Israeli men and women are more depressed than men and women in all the other SHARE countries. Israeli men also report more long term problems. In regard to body mass index, Israeli women are among the most overweight of the women in the SHARE countries, but Israeli men are among the least overweight. In terms of physical activity, Israeli women and men are among the least active in all the SHARE countries. In fact, Israeli women are the least active among their SHARE counterparts, except for Italy. Israeli men are even worse off. They constitute the most inactive men in the SHARE countries.

Comparisons of work status reveal that a relatively high proportion of Israeli women are employed, and a lower proportion of Israeli men are retired than in the corresponding SHARE countries. This may stem from the fact that the age distribution in Israel is younger, and there are relatively more young-olds in Israel than in the European countries. However, Israeli men also reported a higher rate of disability, a status not necessarily affected by age distribution. Another possible explanation for the different rates of retirement is that there are fewer options and less incentive in Israel for early retirement than there are in many European countries.

The last trend of note relates to gross household income. Comparing the Israeli average and the SHARE average overall, one sees that household income in Israel is lower than in most SHARE countries, but rises relatively when correcting for purchasing power parity. This is particularly evident when comparing median household income. The Israeli median is slightly higher than the SHARE median overall, and higher than the country level median in Austria, Italy, Spain and Greece. However, after adjusting for PPP the Israel median rises above the midpoint of the SHARE countries.

These trends point to several areas that will require attention in the formulation of public policy on behalf of the aging population in Israel. The data from SHARE-Israel should also provide many additional opportunities to examine the conditions of the older population in Israel in relation to similar cohorts in Europe and in America. As such, it should constitute an important resource for gerontological research and for social policy development.

Conclusion

This article is the first to report on the findings of the SHARE-Israel project, the first HRS-inspired survey of persons age 50 and over to be executed in the Middle East. While the findings are drawn from a partial preliminary sample only, they nevertheless shed light on the uniqueness of the aging experience within the State of Israel. In some respects, the veteran Jewish-Israeli population is similar to other older populations in Europe, but in other respects it is different. The Arab-Israeli population, on the other hand, differs quite significantly from European older adults across many parameters. However, it is the older immigrants from the former Soviet Union, a unique group which constitutes a fifth of older Israeli adults, who present the most challenging picture of what it means to be old. The many difficulties that emerge among this population group suggest that late-life immigration carries numerous risks that need to be addressed by public policy.