CI: Confidence Interval; SD: Standard Deviation; ETB: Ethiopian Birr; TFR: Total Fertility Rate; EDHS 2011: Ethiopian Demographic and Health Survey 2011; Kedge: Knowledge; HH: Household; Vs: Verses; Gov: Government

Factors affecting family size contribute to fertility rate. Rapid growth of population is obstacle to harmonize the rate of population growth with socio-economic development. Rapid growth is occurring in countries where the decline in total fertility rate is slow. The slow decline of total fertility rate may be because population has not reached their desired family size. The high and nearly stable desired family size is an obstacle to further fertility decline. However, little is known about the determinants of desired family size in urban and less urban area. This study determines factors influencing family size preference among residents of Assela town so as to develop ways to handle rapid population growth in developing countries.

The continuing growth of the world population has become an urgent global problem. Most of this growth is occurring in developing countries where the decline in total fertility rate is slow. The slow decline of total fertility rate may be because population has not reached their desired family size [1,2]. Variables reflecting desired family size are strong predictors of the numbers of children born to women [3]. Family size denotes the total number of children a woman has borne at a point in time [4]. Family size depends on numerous factors, such as age, duration of marriage, literacy, preference of number of children, etc. [5]. In most Sub-Saharan countries in which fertility has declined, desired family size is lower and demand for contraception is relatively high [6].

Researchers have neglected to investigate “population’s desire” for a particular family size and policy makers have failed to design careful and targeted communication in this regard [7]. Although the information pertaining to desired family size has been gathered in fertility related surveys in some developing countries like Pakistan, an assessment of these variables as the factors influencing the achieved family size remains problematic [8].

Without the knowledge of factors affecting family size, the desire to achieve small family size and harmonize the rate of population growth with socioeconomic development in order to reach a high level of welfare cannot be achieved [9].

Trends in family size preferences have important implications for trends in fertility [10]. The high and nearly stable desired family size is an obstacle to further fertility decline [11]. Pessimism has been expressed about the trend of family sizes and its tendency for a probable world population explosion which could plunge poor developing countries into further poverty [2].

There is a strong inverse correlation between these variables[desired family size and contraception: the lower the desired family size, the larger the demand for contraception, which is as expected because contraception is the main means by which women implement their preferences for smaller families [12].

With an estimated population of 77 million, Ethiopia is the second most populous country in Africa next to Nigeria [13]. Different factors influence family size at different level in different societies [14]. This study assessed factors influencing desired family size in Assela town where data was not available in this regard.

Study design, setting and participants

This study is a community based cross-sectional study conducted in Assela town, which is located at 175 kilometers to the South-East of Addis Ababa. The town has 14 kebeles. The 2007 National Census reported a total population for the town was 67,269, of whom 33,826 were men and 33,443 were women. The majority of the inhabitants (67%) practiced Ethiopian Orthodox Christianity, while 22% of the populations were Muslim, and 8% of the populations were Protestants. Assela town has one public hospital, one health centre, five private clinics and four NGOs working on reproductive health. The study was conducted from March 25 to April 4, 2013 in 5 kebeles of Assela town, Ethiopia.

Sample size and sampling procedures

The sample size was determined by using a formula for estimation of single population proportion with the assumption of 95% confidence interval, 5% margin of error and prevalence of 50%. To compensate for the non-response rate, 11% of the determined sample was added up on the calculated sample size and the final sample size was 428. The study units were households, assuming that each household would have married couples. Information was obtained from each kebele office regarding the person who can able to find and obtain the households. A guide, who knows the selected Kebele very well, was recruited for each selected Kebele. The borders at entrance of the selected Kebeles were used as starting point for the random selection of the households. The number of intervals from one household to another household was already identified by dividing the total number of households in the Kebele to the allocated sample size. Using the Guide, from the border which identified the kebeles the nearby houses relatively straight to each other were counted until the number of household in one interval were attained. Out of the households counted within the interval, one household was selected by lottery method. That identified household was used as the first household for the study. The subsequent households were identified by the interval calculated for each Kebele.

Data collection procedures

Structured questionnaire for interview was developed in English and translated into Afaan Oromo language version for data collection. The questionnaire contains general information including Sociodemographic characteristics of the residents and their family as well as the question which assesses factors influencing family size preferences. The questionnaires used for collecting information from both men and women were similar.

The data was collected by grade 12 complete five males and five females. The data collectors were supervised by Bachelor of Science holders in health. The data collector and supervisors were trained for two days.

The training consisted of the objectives of the study, introduction of questionnaire format, procedure of interviewing the respondents and method of reporting to immediate supervisor. The communication between supervisor’s, data collectors and investigator was thoroughly explained. Trained data collectors collected data on the variables from couple’s over 9 days. When the couples couldn’t be found after two visits, the data collectors chose the next household in either side of the visited one. The data collectors were paired in such a way that a pair consisted of a male and a female.

During the process of data collection, the male data collector interviewed the husband while the female data collector interviewed the wife at the same time, but in separate location at a distance that the couples couldn't hear each other. That was to avoid the probable influence of one partner on the other partner on the information to be gathered. It is hoped that female data collectors could best able to solicit open and honest replies from the female respondents on sensitive issues. A supervisor was responsible for collected data, checking for inconsistencies and omissions. Submission of filled formats to coordinator was made every day. Formats with problem were sent back to supervisors for re-interview.

Data processing and analysis

The collected data was entered into computer for analysis by using statistical packages: EPI-INFO version 3.5.1. The data was exported to SPSS version 16 for analysis. Simple frequencies to see the overall distribution of the study subject with the variable understudy was done.

The outcome variable that is desired family size was measured using mean score. Respondents who had no children were asked how many children they would like to have if they could choose the number of children to have over their entire lifetime. Those who had living children were asked the number of children they would choose if they could start their childbearing again. And it was analyzed using mean score.

Independent variables included were household income, religion, education, mother’s health, psychological factors, and knowledge of contraception, child sex preference, age at marriage and other factors. Thus, respondents were asked to determine which factors influenced their preferences on the size of their family and were documented as either ‘’yes” or “no”. The proportion of respondents who were answered “yes” for each factor was defined as magnitude of the factor on family size preferences. All respondents were asked to rank those identified factors according to the importance for their preferences. The most important one was placed as the first rank and the least important rank was last. Multiple linear regressions were used to see relationship between influential factors and desired family size. Significance level and association of variables were tested by using 95% confidence interval (C.I) and regression coefficient. The findings were presented by text and tables.

Ethical consideration

Ethical Clearance was obtained from the Research and Ethics Committee of School of Public Health, College of Health Science, Addis Ababa University. Additionally an informed verbal consent was obtained from each respondent after providing sufficient information on the purpose of the study. To ensure the confidentiality of respondents their names were not written on the questionnaire. All interviews were made individually to keep privacy.

Over all description of study population

A total of 428 residents were included in the study. The mean age of the women was 30 years (SD=6.95). The mean age of men was 37 years (SD=12.92). The mean age difference was 6.44 which was statistically significant different (p<0.001). The median ages at first marriage were 19 and 22 for women age 15-49 and men age 15 or above years respectively. Men tend to marry older than women. The minimum age at first marriage for women was 15 years.

Most of the respondents were Orthodox 229 (54%) with the remaining being Muslims 102 (23.9%), Protestants 84 (19.7%) and the rest are other religious groups. Majority of the population were Oromo 278 (65.45) followed by Amhara 102 (24.2%), Gurage 25 (5.9%) and the rest are other ethnic groups. The distribution of education among men and women were different and 67 (31.5%) of men 51 (24.1%) of women had completed primary school. Most of the women 140 (65.4%) had no their own income. Almost all men 197 (92.1%) were employed, of which 140 (68%) had monthly income of less than 1000 ETH Birr (Table 1).

Table 1: Socio-demographic characteristics of couples residing in Assela town, South-East Ethiopia, March 25 to April 4, 2013 (*Monthly income is in Ethiopian birr, one US $=23 Ethiopian birr).

Description of desired family size by socio-demographic characteristics

The mean desired family size for all study population was 3.8 (SD=2.36). The mean desired family size for men and women were 4.1 (SD=2.05) and 3.5 (SD=2.60) respectively. The mean difference in the number of children desired by men and women was 0.6 which is significant difference (p=0.004). Men preferred larger family size than women. There are variations in the mean ideal number of children by background characteristics among all men and women. The older the respondents, the more children that they desired; women age 15-19 responded that the mean ideal family size is 2.5 (SD=1.50) children, while women age 35+ said 4.6 (SD=2.31).

Women belonging to Protestant and catholic religious groups were found to have relatively lower mean desired family size when compared to those following Orthodox Christian and Muslim. The mean ideal number of children declined as levels of education of both women and men increased with the exception of greater than secondary level.

For example, the mean ideal number of children among women who had completed primary school was 4.1 (SD=1.18), compared with 5.8 (SD=1.00) children among women with no education. The mean numbers of children desired by couples of different educational levels are significantly different (p<0.001) (Table 2).

Table 2: Socio-demographic characteristics and mean desired family size of couples residing in Assela town, South-East Ethiopia, March 25 to April 4, 2013 (*Monthly income is in Ethiopian birr, one US $=23 Ethiopian birr).

Magnitudes and ranks of the factors influencing desired family size

Factors influencing desired family size which were identified included house hold income, sex preference, psychological, mothers health, education, religion, knowledge of contraception and age at marriage. The couples were asked to determine whether these factors were influenced their desired family size or not and to rank them. The response on magnitudes and ranks of the factors are shown below (Table 3).

Table 3: Ranks and proportions of factors that influenced desired family size of couples residing in Assela town, South-East Ethiopia, March 25 to April 4, 2013.

Association between gender, knowledge of contraception educational status and desired family size

Ten questions for each of the contraceptive methods were developed. Awareness of the respondents on contraception was measured. Then, knowledge score was computed based on the awareness questions. The score is calculated by adding values given to each of the ten questions. Then, it was categorized as poor, moderate and good knowledge.

Those who knew 0-3 correct answers from ten knowledge questions were categorized as “poor knowledge”, while those who knew 4-6 and 7-10 correct answers from ten knowledge questions were categorized as “Moderate knowledge” and “good knowledge” respectively. Based on the composite knowledge score, 165 (40.2%), 160 (38.9%) and 86 (20.9%) of respondents had poor, moderate and good knowledge fertility awareness method of contraception respectively. Women are 2.4 times more likely to prefer small family size compared to men (AOR=2.47 95% CI=1.48-4.11).

Moreover, women who attained secondary and higher level of education were found to be 2 and 2.6 times more likely to prefer small family size compared to women who had no education, respectively (AOR=2.10, 95% CI=1.11-3.98) and (AOR=2.60, 95% CI=1.15-6.77). Women who had good knowledge of fertility awareness method of contraception were found to be 1.7 times more likely to prefer small family size compared to women who had poor knowledge of fertility awareness method of contraception (AOR=1.71, 95% CI=1.08-3.72) (Table 4).

Table 4: Logistic regression showing association between gender, knowledge of contraception, educational status and desired family size of couples residing in Assela town, South-East Ethiopia, March 25 to April 4, 2013 (*Statistically significant at: P-value<0.05).

Association between selected socio-demographic variables and desired family size

Statistically adjusting for the effects of socio-demographic variables listed in the Table 5 age at marriage was negatively associated with desired family size evidencing that as the age at marriage of the respondents increased the desired family size decreased (p<0.001).

Table 5: Results of multiple linear regressions showing the associations between selected Sociodemographic variables and desired family size of couples residing in five kebeles of Assela town, South-East Ethiopia, March 25 to April 4, 2013 (®=reference category *P<0.05, **P<0.01, ***P<0.001, C.I=Confidence Interval, B=Coefficient of regression).

Statistically adjusting for the effects of socio-demographic variables listed in the Table 5 respondents who had no education desired higher family size than respondents who had more than secondary education (p<0.001) (Table 5).

The mean desired family size for all study population was found to be 3.8 in this study. The mean desired family size for men and women were 4.1 and 3.5 respectively. Our finding for women is significantly lower than the finding for women in urban Ethiopia which was 3.7 and for women in Oromia which was 4.3 reported by Demographic and Health Survey of Ethiopia 2011 [15].

The difference could be explained by the existed evidence that is there are variations in mean desired size of family among studies done at different places or in different societies, or at different time and with different sample size or methodology. In Ethiopian Demographic and Health Survey which were done within 5-7 years interval variations the mean preferred family size were; 4.1 in 2000, 3.7 in 2005 and 3.7 in 2011 for women living in urban. Other study which used data from Ethiopian Demographic and Health Survey 2005 showed that the mean number of children desired by women of 15 to 49 years age group was 3.12 [16]. These differences could be due to differences in time the studies were done (Fertility preferences may vary with time) [9], differences in societies and methodology.

A survey which aimed at fertility transition in Rwanda found that the ideal family sizes were 4.1 in 2005 and 2.5 in 2010 among all Rwandan women age 15 to 49 [17]. The mean desired family size that was found for women by our study is lower than the finding of the above survey done in Rwanda in 2010. However, the range of variations in mean desired family sizes since 2005 to 2011 is 2.5 to 4.5 among the studies listed above which were done in developing countries like Ethiopia, and our finding is within this range.

Other similar findings include; a study which used DHS surveys done in the last 15 years (1998 to 2007) from 30 sub-Saharan African countries demonstrated that differences in desired family size among countries are large, with the following low-high ranges of desired family size (births per woman): from 2.3 in Swaziland to 8.6 in Chad [18].

In our study there are variations in the mean ideal number of children by age and sex among all study population. The mean desired family size for men was larger than women’s and the older the respondents the more children that they desired. Women age 15-19 responded that the ideal family size is 2.6 children, while women age 45 to 49 said it is 5.6. Similarly study which used data from Ethiopian Demographic and health Survey to assess the factors influencing women’s intention to limit child bearing in Oromia in 2009 reported that mean desired family size was highest among women aged 40-49 years [19].

Older women may inflate their stated desired family size to be in accord with the number of children they already have, which may include unwanted births (“rationalization”). To minimize this potential error, the mean desired family size among women aged 20 to 29 are being in use as the indicator of fertility preferences. So let’s compare the findings of mean desired family size for this age group with other studies. In our study the mean desired family size among women aged 20 to 29 was 2.95. However a study which used Ethiopian Demographic and Health Survey 2005 showed that mean desired family size among urban women aged 20 to 29 to be 3.37 [16]. The difference could be explained by the time the studies were done.

Factors influencing desired family size which were identified included house hold income, sex preference, knowledge of contraception, education, age at marriage, religion, mothers health and psychological factors. Consistent to our finding study done in Nigeria in 2010 revealed that income level, sex preference, religion, educational level and knowledge of contraceptive methods were significantly associated with family size preferences (p<0.05) [20].

Other studies with similar findings include: A study done in Pakistan in 2007 showed that sex preference, family income and age at marriage influenced family size [21]. Similarly study in Bangladesh in 2011 showed that religion, education, age at marriage and son preference were related to family size [22]. However, not all the factors found by the above studies are similar including our study. For example; in the above study done in Bangladesh in 2011 household income and knowledge of contraception was not found to be related to family size but these factors were found to be related to family size in our study and in the study done in Nigeria in 2010. This could be due to the fact that different factors influence family size in different societies [14].

Regarding the magnitudes of factors influencing desired family size, our study found that household income, mother’s health, sex preference and psychological factors had got 65.4%, 25.5%, 17.8% and 22.9% respectively. However, household survey done in Ekpoma (Ekpoma is a growing University town situated in Esan West Local government Area of Edo State), Nigeria in 2007 found that about 94% of the study population insisted on the sex of the children as the main factor. Only 4% believe economic consideration was the factor responsible for the size of the family [23]. The above differences in magnitude of the influential factors between our study and the above survey done in Nigeria could be explained by fact that different factors influence family size at different level in different societies [14]. That is there are variations in magnitude of the factors influencing family size preference among studies done in different societies or at different time.

The magnitude of sex preference and psychological factors in our study was 17.8% and 22.9% respectively. However a survey conducted among 200 women in Pakistan during the period of September to December 2009 showed, 19.5% preferred particular family sizes because their family requires sex composition and 42% of women answered they like the size without any reason [12]. This finding is not only different from the finding of our study but also it is different from the findings of other studies listed above which could be due to the fact that the studies were done in different societies.

In this study, women prefer small family size 2.2 times more likely compared to men, similarly study done in Egypt in 2011 found that, young females in households with one or more adults preferring a small family are 78 percent more likely to desire a small family, and young females in households with one or more young people who prefer a small family are 37 percent more likely to desire a small family themselves, compared with those living with adults or with young people, respectively, who do not prefer a small family [24]. In this study, women who attained secondary and higher level of education were found to be 2 and 2.6 times more likely to prefer small family size compared to women who had no education similarly, study done in Ethiopia in 2007 showed the importance of variables such as maternal education for smaller family size and desire on the part of women to limit family size, especially as the number of surviving children increased [3].

The directions of the effect of the factors influencing desired family size in our study were similar with most of the studies done in developing countries cited above. In our study sex preference and psychological factors forced the couples to have larger family size than other factors. Similarly in a study done in Pakistan in 2011 son preference and lack of consensus between husband and wife on number of children forced couples to have large family size or more than two children [9]. In our study desired family size decreased as the age at marriage increased. Similarly study done in Bangladesh in 2011 showed that desired family size decreased as the age at marriage increased [22]. In our study women with secondary or higher education have, on average, lower desired family size than women with no education (2.9 vs. 5.8 births per woman). Similarly a survey which used DHS surveys done in the last 15 years (1998 to 2007) from 30 sub-Saharan African countries demonstrated that women with secondary or higher education have, on average, lower desired family size than women with no education (3.7 vs. 5.6 births per woman) [18].

The mean desired family size for the study population was 3.8.

The mean desired family size for men and women were 4.1 and 3.5 respectively.

Respondents with no education desired higher family size than respondents who had more than secondary education.

Respondents with primary education desired higher family size than respondents who had more than secondary education.

As age at marriage of study population increased mean desired family size decreased.

It is quite possible that increasing educational level and age at marriage might influence couples to desire lower family size.

Local government should focus on increasing educational level of the community.

Programs of family planning services should aim to reduce fertility rates by focusing not only on expanding contraceptive prevalence but also on creating awareness about factors influencing family size. Qualitative methods should be used to identify other factors influencing desired family size in addition to quantitative research.

My deepest thanks go to my advisor, Dr. Getnet Mitike for his beneficial assistance and critique starting from the development of the proposal up to the end of the research project.

At last but not least my thanks go to the study participants, who devoted their time to give valuable information.

Bekele Dibaba carried out the research from conception to the write up of the final article.

Getinet Mitike was participated in developing the proposal and adviser of the research project.