CHAPTER 4 Mixed Evidence of the SDT in Taiwan
4.1 Changes Over Time
4.1.1 Profile of the Survey Respondents
國
立 政 治 大 學
‧
N a
tio na
l C h engchi U ni ve rs it y
47
survey. Discussion of these findings, combined with discussion of the findings of the qualitative analysis of interviews with Taiwanese women presented in the next chapter, are presented in Chapter 6.
4.1 Changes Over Time
This section looks at changes in the composition of the sample profiles and the value orientations of the overall samples of Taiwan Social Change Survey general questionnaires from Round 1 and 7. This 30-year time period is significant because the Round 1 survey was conducted at the same time that fertility dropped below replacement levels in Taiwan (1984/1985), an essential indication of demographic change associated with the Second Demographic Transition. Round 7 (2015) is the most recent round conducted and represents Taiwanese society today, three decades after the demographic changes that potentially signal the SDT began.
4.1.1 Profile of the Survey Respondents
The profile of the general survey samples for Round 1 and 7 reflect the some of the societal changes that occurred during those thirty years. For details about the composition of each survey sample by age, gender, marital status, education level, and income, see Table 4.1. The full 2015 sample was limited to age seventy-five and under in all of the following comparisons in this section to match the age range included in the 1985 sample. In 1985, the age group that consisted of the largest percentage of the sample was 30-39 years old at 29 percent, but in 2015 the age groups were roughly equally distributed, ranging from 19 to 21 percent.
‧
Table 4.1: General Survey Sample Profiles by Age, Gender, Marital Status, Education Level, and Income, Rounds 1 and 7
Age Group Percentage by Year
Under 30 30-39 40-49 50-59 60 and over
Gender Percentage by Year
Male Female
1985 2015
62 38 n=4,199
52 48 n=1,877
Marital Status Percentage by Year
Married Formerly
Married Single Cohabitating
1985 2015
80 7 13 n/a n=4,192
57 11 31 <1 n=1,877
Education Level Percentage by Year Junior
Monthly Household Income Level Percentage By Year Under
Note: Due to rounding, percentages may not add up to 100
The gender ratio between samples was radically different. Men made up 62 percent of the sample in 1985 and 52 percent of the sample in 2015. These percentages do not reflect the actual gender ratios in either sample year. In 1981 there were about 1.09 males for every female but the sample from 1985 has about 1.6 males for every female.
In 2015 there were about 0.99 males for every female, but in the 2015 sample there are
‧
國立 政 治 大 學
‧
N a
tio na
l C h engchi U ni ve rs it y
49
1.04 males for every female (DGBAS, 2016). In both cases, there was a disproportionate number of males included in the sample.
The composition of marital status also reflects substantial changes between the two rounds, as expected due to the changes in marriage behavior that occurred during that period. In 1985 by far the greatest percentage of the sample was married, at 80 percent.
In 2015 the percentage of respondents married was still the largest percentage, but dropped to 57 percent. The percentage of respondents formerly married increased from seven to 11 percent and the percentage of single respondents more than doubled,
changing from 13 percent to 31 percent. 2015 had a “cohabitating” option, but less than one percent of the sample chose it. The real rates of cohabitation are likely higher but stigma may have prevented respondents from reporting it.
Samples from Round 1 and Round 7 also show a shift toward higher levels of education, which was expected. In 1985, the largest percentage of respondents had attended junior high or below, at 67 percent, with the next largest percentage high school level, at 19 percent. In 2015, those with junior high education or lower dropped to 28 percent, which was equal to the percentage of respondents with high school level
educations and similar to the percentage with university educations. The percentages of those with a university level education more than tripled, rising from seven percent to 26 percent, while percentages of respondents attending graduate school or higher increased from two to seven percent.
Monthly household income levels cannot be compared between the two rounds because the general survey in Round 1 did not include a question about income. In 2015 the income distribution is also unclear because 24 percent of the sample reported “other.”
Looking closer at the respondents that reported “other,” 60 percent were either under 30 years old or were 60 years old and over, so it is possible most of those respondents were being supported by family and were uncertain of the household income. Aside from
“other,” the most common income level was 30,000-59,999 at 21 percent.
‧
國立 政 治 大 學
‧
N a
tio na
l C h engchi U ni ve rs it y
50 4.1.2 Changes in Overall Value Orientation
As can be seen in Table 4.2, there was a clear shift in overall value orientation between survey rounds. In both rounds, the largest proportion of respondents were neutral, but there was a drop from 70 percent in 1985 to 54 percent in 2015. The number of respondents with moderately conformist and conformist value orientations dropped from 27 percent to four percent while the number of respondents with moderately nonconformist and nonconformist value orientations increased from three percent to 42 percent.
Looking at the breakdown of values each year by age, also shown in Table 4.2, these shifts were not driven by a difference in values among different age cohorts. In both survey years, the distribution of values is very similar across age groups. In 1985, respondents with conformist value orientations were less than one percent for each age group. Respondents with moderately conformist value orientations ranged from 25 percent to 32 percent, neutral orientations ranged from 63 percent to 72 percent, and moderately nonconformist orientations ranged from two to five percent for each age group. None of the respondents in 1985 had nonconformist value orientations. In 2015, none of the respondents had conformist value orientations, moderately conformist ranged from two to eight percent, neutral ranged from 49 to 65 percent, moderately
nonconformist ranged from 27 to 48 percent, and nonconformist ranged from zero to less than one percent for each age group. The 60 and over age group in 2015 does show higher percentages of moderately conformist and neutral orientations and lower
percentages of moderately nonconformist orientations compared to those below 60, but for all age groups the largest proportions of respondents were neutral, followed by moderately nonconformist, moderately conformist, and nonconformist, in that order.
‧
Table 4.2: Overall Value Orientation Percentages and Value Orientation Percentages by Age, Gender, Education, and Income for General Surveys, Round 1 (1985) and Round 7 (2015)
Overall Value Orientation Percentages
Conformist Moderately
Overall Family Value Orientation Percentages by Age (Column Percentage)
Under 30 30-39 40-49 50-59 60 and Over
Overall Family Value Orientation Percentages by Gender (Column Percentage)
Male Female
‧
Overall Family Value Orientation Percentages by Education (Column Percentage) Junior High
or Below
High School Junior College
Overall Family Value Orientation Percentages by Monthly Household Income (Column Percentage)
2015 Only Under 30,000 30,000-59,999
Note: Due to rounding, percentages may not add up to 100
Breaking down value orientation by gender also shows very similar distributions for men and women. As shown in Table 4.2, in 1985 respondents with conformist value orientations were less than one percent, moderately conformist ranged from 26 to 28 percent, neutral ranged from 68 to 71 percent, and moderately nonconformist was three percent for each gender. In 2015, respondents with moderately conformist value orientations ranged from three to four percent, neutral ranged from 52 to 56 percent,
‧
國立 政 治 大 學
‧
N a
tio na
l C h engchi U ni ve rs it y
53
moderately nonconformist ranged from 40 to 43 percent, and nonconformist was less than one percent.
Unlike age and gender, value orientation does change across the different education levels for both survey rounds. In 1985, all of those with conformist value orientations had junior high school level education or lower. Respondents with junior high or lower levels also had a higher percentage with moderately conformist orientations compared to other education levels. For respondents with high school education and above, the distribution of values across levels was similar. Between 14 and 19 percent had moderately conformist orientations, between 73 and 81 percent had neutral
orientations, and between five and nine percent had moderately nonconformist value orientations.
In 2015, the differences in values across education levels are more striking. The percentage of respondents with moderately conformist value orientations was eight percent for those with junior high education level or lower and decreased with each increase in education level. For respondents with graduate school or higher, less than one percent had a moderately conformist value orientation. Junior high and below education level also had the highest percentage of respondents with neutral orientations at 70 percent. As education level increased, the percentage of respondents with neutral value orientations at each level decreased. For respondents with graduate school or higher, 26 percent had neutral value orientations. The percentage of respondents with moderately nonconformist orientations was highest for those with graduate school and above
education level at 72 percent and decreased with each decrease in education level. Only 22 percent of respondents with junior high and below education had moderately
conformist orientations. Finally, only respondents with university education and above had nonconformist value orientations. Overall, it is clear that in the 2015 sample respondents with lower levels of education had higher proportions of moderately conformist and neutral value orientations while respondents with higher levels of
education had higher proportions of moderately nonconformist and nonconformist value orientations. There are also bigger differences in values across education levels in 2015 than in 1985.
‧
國立 政 治 大 學
‧
N a
tio na
l C h engchi U ni ve rs it y
54
The breakdown of value orientation by income levels cannot be compared between rounds because the 1985 survey did not include a question about income. For 2015 the relationship is somewhat unclear because about a quarter of respondents chose the “other” category. Looking at the ages of respondents who chose “other,” the majority were under 30, 21 percent, or over 60, 39 percent, so it is possible that these respondents had low income or relied on others for their livelihood. This is further supported by the distribution of value orientations for respondents who chose “other” because it is nearly identical to the distribution for respondents with under 30,000 NTD/month. Aside from the “other” category, proportions of respondents with moderately conformist orientations were highest at lower income levels and ranged from six percent to one percent. Higher percentages of respondents also had neutral value orientations at the lower income levels and decreased as income increased. Neutral orientations ranged from 65 percent to 43 percent. Moderately nonconformist proportions rose as income increased, from 30 percent at the lowest income level to 55 percent at the highest level. Nonconformist value orientation made up less than one percent or one percent of respondents at each income level from 60,000 NTD/month and up.
In summary, overall value orientation shifted from mostly moderately conformist and neutral in 1985 to mostly neutral and moderately nonconformist in 2015. This
finding supports the expectation of value shift in Second Demographic Transition theory.
Surprisingly, in 1985 the distribution of value orientations across age groups was very similar. In 2015, the distribution of value orientations was very similar for different age groups of respondents under 60, while respondents over 60 had slightly higher
percentages of moderately conformist and neutral value orientations and much lower percentages of moderately nonconformist value orientations. In both survey rounds, the distribution of values for men and women was very similar.
Values did vary by education in both rounds and income in 2015. In 1985 there was a divide between those with junior high and below education levels and those with higher education levels, the former having higher proportions of conformist and
moderately conformist value orientations and the latter having higher proportions of neutral and moderately nonconformist orientations. In 2015, proportions of moderately
‧
國立 政 治 大 學
‧
N a
tio na
l C h engchi U ni ve rs it y
55
conformist and neutral respondents decreased as education level increased and
proportions of moderately nonconformist and nonconformist respondents increased as education level increased. Finally, in 2015 respondents with lower incomes had higher proportions of moderately conformist and neutral orientations while respondents with higher incomes had higher proportions of moderately nonconformist and nonconformist value orientations. The next section looks at changes between each category of values between survey rounds.
4.1.3 Changes in Religious Attitude, Attitude toward Education, and Family Value Orientations
In each general survey round, there were three categories of values that were analyzed: religious attitude, attitude toward education, and family values (called “moral concepts” in the questionnaires). This section will first compare the changes over time for each value category and then compare the difference between family and nonfamily values for each round.
The distribution of overall religious attitude for each survey round is very similar between survey rounds, though there are differences for individual questions. Questions included belief in a spirit’s existence after death, the importance of ancestral worship, and reliance on the gods versus one’s own hard work. From 1985 to 2015 the percentage of respondents with conformist religious value orientations increased very slightly from just under one percent to about one and a half percent. Moderately conformist orientations also barely increased from 19 percent to 20 percent. Neutral orientations decreased from 66 percent to 63 percent. Moderately nonconformist orientations increased from 13 percent to 15 percent. Nonconformist orientations were about one percent for both years.
As can be seen in Table 4.3 the overall changes in religious attitudes were very small.
‧
Table 4.3: Value Percentages by Religious Attitude, Attitude toward Education, and Family Values for General Surveys, Round 1 (1985) and Round 7 (2015)
Conformist Mod.
Note: Due to rounding percentages may not add up to 100; for 1985 N=4,199; for 2015 N=1,877
‧
國立 政 治 大 學
‧
N a
tio na
l C h engchi U ni ve rs it y
57
Looking at the individual questions, from 1985 to 2015, belief in a spirit
continuing after death actually increased, from 53 percent to 69 percent, while belief that it is important to have descendants offer ancestor worship for one after death declined from 85 percent to 67 percent. In both survey rounds the vast majority of respondents believed that one did not need to rely on the gods if they worked hard. In 1985 86 percent believed they could rely on hard work and in 2015 81 percent did. The
percentage of those who believed they still needed to rely on the gods increased from ten percent to 17 percent. Increases in religious beliefs related to a spirit after death and reliance on the gods is likely related to increased percentages of respondents reporting a religious affiliation. In 1985, 68 percent of respondents reported “none” for religious affiliation, which dropped to 21 percent in 2015. Increases were not due to growth in Western religions, as Catholicism, Protestantism, and Islam accounted for seven percent or less in both rounds. Rather, respondents reporting affiliation with Buddhism, Taoism, and Folk Religion increased from 25 percent to 71 percent. Belief in the importance of ancestor worship declined despite this increase.
The second value category, attitude toward education, shows moderate changes over time. However, this should be interpreted with some caution because the
questionnaire options were slightly different between rounds. For the question regarding purpose of education in 1985, the most popular answer was “to make life more fun and interesting,” which was coded as moderately nonconformist. This answer didn’t exist in the 2015 survey and though all other choices were identical, that option may skew 1985 education value orientations toward the nonconformist side. For the questions regarding the lowest level of education acceptable for boys and girls to attain, the most popular answer for boy’s education and second most popular answer for girl’s education was “as high as possible.” This answer was coded as neutral because it is unclear exactly what level this means since it would be different for each respondent. As this answer also did not exist in the 2015 survey though all other choices remained the same, these two questions might have skewed the 1985 overall attitude toward education toward neutral or moderately nonconformist.
‧
國立 政 治 大 學
‧
N a
tio na
l C h engchi U ni ve rs it y
58
Even taking into consideration that the 1985 results for overall education value orientation might be skewed compared to if the answer choices were identical to the 2015 questionnaire, there is still a shift toward the nonconformist side. For both rounds, conformist orientations are less than one percent. Moderately conformist orientations dropped from 15 percent to six percent. Neutral orientations dropped slightly from 57 to 52 percent. Moderately nonconformist orientations increased from 28 to 40 percent and nonconformist orientations increased from less than one percent to one percent.
There is also a clear shift toward preference for higher education levels. In 1985, the second most common answer after “as high as possible” for the lowest level of education a boy should have was “high school.” For the lowest level of education a girl should have, “high school” was the most common answer, followed by “as high as possible.” In 2015, the most common answer for both boys and girls was “university,”
with over 50 percent of respondents choosing that option.
The final value category is related to family, though it was labelled in the
questionnaires as “moral concepts.” Questions ask about how wrong it is for a daughter-in-law to argue with her mother-daughter-in-law due to differences in opinion, to send one’s parents to a nursing home, and to get a divorce because the couple cannot get along. Out of the three value categories, family values showed the most dramatic shift. For overall family value orientations from 1985 to 2015, conformist decreased from 25 percent to two percent, moderately conformist decreased from 44 percent to nine percent, neutral decreased from 26 percent to 20 percent, moderately nonconformist increased from five percent to 40 percent, and nonconformist increased from less than one percent to 30 percent. It is clear that family values changed considerably over the 30 years between survey rounds.
Looking closer at each family question, values regarding obedience of a daughter-in-law to her mother-daughter-in-law show a definite shift. In 1985 64 percent of respondents were either conformist or moderately conformist versus 75 percent of respondents being nonconformist or moderately nonconformist in 2015. Values about caring for elderly parents in 1985 showed that an overwhelming number of respondents thought it was wrong to put parents in nursing homes. 91 percent were either conformist or moderately
‧
國立 政 治 大 學
‧
N a
tio na
l C h engchi U ni ve rs it y
59
conformist for that question. In 2015, there was a shift toward greater acceptance of putting parents in nursing homes, though about a third still felt it was very wrong. 33 percent were either conformist or moderately conformist while 59 percent were either nonconformist or moderately nonconformist. Finally, acceptance of divorce before children were fully grown increased drastically between survey rounds. In 1985, 56 percent were either conformist or moderately conformist but in 2015 83 percent were nonconformist or moderately nonconformist.
After breaking down value change by value categories, it is clear that the majority of overall value shift was due to changes in family values. Religion showed very little overall change and education showed only moderate change. It’s possible attitudes toward education shifted more than these results reflect, especially since level of education increased between survey rounds, but to what degree is unknown from this data.
4.2 Value Orientation and Household Type in 1985 and 2015
This section explores the relationship between value orientation and household type in 1985 and 2015. First, a regression analysis was conducted on data for each year using overall value orientation as the explanatory variable. The findings for these regression models are described in section 4.2.1. Second, an additional regression analysis was run on data for each year using two explanatory value variables, family
This section explores the relationship between value orientation and household type in 1985 and 2015. First, a regression analysis was conducted on data for each year using overall value orientation as the explanatory variable. The findings for these regression models are described in section 4.2.1. Second, an additional regression analysis was run on data for each year using two explanatory value variables, family