Studentnummer: 0181245



Student Registration Number:

Term paper

- The Silent Revolution

A test of the Inglehart hypothesis -

Hand-in date:

08.06.2009

Campus:

BI Oslo

Exam code and name:

GRA 5917 Public Opinion and Input Politics

Program:

Master of Science in Leadership Political economy

Content

Content i

List of figures i

Summary ii

Summary ii

Introduction 1

Research topic 2

Causality 2

Academic contribution 3

Theory 3

Data 5

Independent variables 5

Dependent variables 6

Methods 6

Analysis 7

Conclusions 13

Bibliography 15

List of figures

Figure 1 Maslow’s hierarchy of needs 4

Figure 2 Scatter dot diagram, Postmaterialism and GDP 7

Figure 3 Correlation matrix 8

Figure 4 Descriptive statistics 9

Figure 5. Linear regression analysis. 10

Figure 6. Binary logistic regression 11

Figure 7 Fixed effects models 12

Summary

Ronald Inglehart (1977) claims that postmaterialism in a society is a product of increased economic security and security from external threats. His scarcity hypothesis, which is based on Maslow’s hierarchy of needs, says that people will always put greatest value on those goods that are short in supply. Once the basic physiological needs are covered, one will put greater value to other goods not essential for survival. This will eventually lead to a desire to fulfil needs of a non-material nature, like respect of others, morality, lack of prejudice and other values that can be categorised as postmaterial. His second hypothesis is the socialisation hypothesis. Here Inglehart focus on the environment in which people form their values. A person who grows up in a stable wealthy society will have a different set of values than a person who grow up in war struck and poor society.

While Inglehart in his work focus on level of economic development as an important explanation for postmaterialism, he only uses current level of GDP as a measure of this feature. As Inglehart himself says, values are formed in the early stages of life, which is why this article tries to explain postmaterialism by using the level of economic development for each participant of the World Values Survey at the time when they were 20 years old. Through some robust regression techniques can this article conclude that GDP in early stages of life explain as much, if not more, of the variation in postmaterialism as GDP at the time of the survey.

Introduction

The World Values Survey (WVS) is extensive academic project measuring the state of moral, religion, political preferences, trust and several other aspects of life in societies all over the world. The surveys grew out of the European Values surveys in 1981, and have covered an ever increasing number of countries. This is now an extensive dataset which now covers over 80 states, and many of these have taken place in more than one wave of the survey. The number of variables included in the survey is around 3000 and the number of individual responses is around 260.000. This dataset gives social scientists almost unlimited possibilities for quantitative research.

One researcher who has really taken advantage of the possibilities in these data is Ronald Inglehart. Inglehart has published several articles and books based on the findings he has done in the WVS. Among his findings is the shift towards so-called postmaterialistic values in wealthy and stable societies. The concept of postmaterialism will be defined to greater detail later, but in short it implies that people’s preferences shift on questions on what their country’s political aims should be in the years to come. The shift occurs as socio-economic level of the society is getting higher. The postmaterialist-theory suggests that people will to a decreasing degree focus on security and economic as main priorities for their country. And people will to an increasing degree prefer a focus on “issues such as individual freedom, self expression, and participation”(Dalton 2002).

Inglehart (Inglehart 1977; Inglehart 1997; Inglehart and Abramson 1999) have done extensive analysis on the phenomena of postmaterialism. He has investigated changes between regions, changes over time, changes between generations, and much more. What this article will deal with is the causes of postmaterialism. Inglehart explains, to a large extent, postmaterialistic values with the level of economic development in the society. Inglehart’s analysis uses GDP per capita as measure of economic development. That in itself in uncontroversial, but he uses the level of GDP per capita in the year of the survey, and this is where this article will differ from a pure replication of Inglehart’s analysis. In his analysis (Inglehart 1997:151) place the investigated countries in a scatter diagram with level of GNP per capita on the x-axis and postmaterialism on the y-axis. Postmaterialism is here measured by subtracting percentage of materialists in a country from percentage of postmaterialists. This gives each country one value for postmaterialism. My analysis will focus more on the individual level. I will investigate the relationship between individual postmaterialism scores and level of economic development in society for each individual at their formative age. This analysis is important because both Inglehart (1997) and others (2002) argue that values depend on the individuals situation in life at an early stage. But Inglehart still use a non-satisfactory measurement for measuring one of his most important variables.

More generally is postmaterialism argued to be a part of a wider “silent revolution” taking place in the last couple of decades, called post modernism. This revolution has had great impact on party systems, introducing environmental movements and parties. It has also led to a dramatic increase of women in politics. Samuel Huntington (1991) even suggests that it has triggered a “third wave” of democratization. These issues, green politics, equal opportunities, participation, and minority rights have been called “new politics” (Dalton 2002:134), and the increased importance of this dimension can be seen as a result of postmaterialism.

Research topic

This article aims to build on the admirable analysis of Ronald Inglehart but is going to introduce a new explanatory variable. So instead of explaining the variance in postmaterialism by level of economic development at time of survey, this analysis will use the level of economic development at the time in life when people’s values are set. By using this method we will hopefully be able to understand even more of the concept of how postmaterialist values are established in a society.

Inglehart’s theory, that economic security causes postmaterialism, rests on two main hypothesis(Dalton 2002). One scarcity hypothesis, very much related to Maslow’s pyramid of needs. It states that people will value those things that are scarce and tend to take for granted those things that are relatively abundant (Duch and Taylor 1993). The second hypothesis is what Inglehart has called the socialisation hypothesis (Dalton 2002). In my analysis, this will be the most important hypothesis because it suggests “...that values become crystallized early in life; thus, the overall values of a society reflect the conditions decades or more earlier, when values were being formed” (Dalton 2002). So, growing up in an affluent society will increase the probability of holding postmaterialistic values as an adult. This hypothesis will by tested by using historic data of GDP per capita across states of the world.

Causality

Most of the theory on postmaterialism agrees on the causal relationship in this analysis. Increasing postmaterialism is caused by higher levels of economic development. It is likely that there also will be other important variables that explain postmaterialism, like historic and cultural variables. This analysis will be careful to go beyond the work of Inglehart when it comes to choice of variables.

Academic contribution

There is a large literature on postmaterialism and most of it is in some way or another connected to Ronald Inglehart. His first book on the subject was published in 1977 and have over the last three decades published numerous articles, chapters and books in which he expands on his theories of “the silent revolution” (Inglehart 1977), and he makes use of better and more updated data material for each publication.

The Norwegian political scientist Ottar Hellevik has himself done a lot of research on changing values in society, specifically within the Norwegian context (Hellevik and Børretzen 1996). In his book from 1996 he tests Inglehart’s and theories using a data set collected by Norwegian statistics bureau MMI from 1985 – 1995. He argues that his findings from the Norwegian context can not support Inglehart’s theories. In the period he investigates, a period of rapid economic growth in Norway, he can not detect any substantial increase in the share of postmaterialists Norway (Hellevik and Børretzen 1996:84). Except for the first two years of the period, he only finds decline in share of postmaterialists in Norway. Another Norwegian who has done research on this topic is Oddbjørn Knutsen (1979; Knutsen 1989). He started his academic career by testing Inglehart’s theory with data from three small Norwegian industrial towns in his dissertation from 1979. He has continued to work on related themes for the following decades.

Theory

As previously mentioned, Inglehart rests his postmaterialism theory on two main hypotheses. The first one is what he calls the scarcity hypothesis. To a large extent is this in line with the theories of psychologist Abraham Maslow. Maslow created a very famous hierarchy of needs (figure 1), in which he argued that a humans most basic needs, the physiological needs had to be covered before one could move on to cover the needs which were the next in the hierarchy. In his hierarchy, one can see a clear resemblance with what Inglehart describes as materialistic values at the lower steps, and more typical postmaterialistic values, like respect for others and morality, at the higher steps of the pyramid. Inglehart describes the scarcity hypothesis like this: “An individual’s priorities reflect the socioeconomic environment: one places the greatest subjective value on those things that are relatively short in supply” (Inglehart 1997:32). So according to this, persons values will reflect the socioeconomic situation in which the person lives. If for example the water supply is short, then people will put high emphasis on the need for getting clean water every day. But if those basic needs are met, values reflecting these needs are lowered in priority (Knutsen 1989:4).

[pic]

Figure 1 Maslow’s hierarchy of needs

Inglehart’s early work was closely linked to the Maslovian hierarchy, but is less prominent in his more recent work. The second of the main hypothesis are the socialization hypothesis. This thesis suggests that personal values to a greater degree depend on the socialisation process and assume that values and priorities to a large degree are set in the early years of life. According to Knutsen (1989:4) this will suggest a substantial time lag between socio-economic development and change in values in the population. This is somewhat contradictive to the scarcity hypothesis which does not mention any time lag. Inglehart explains that “The relationship between socioeconomic environment and value priorities is not one of immediate adjustment: a substantial time lag is involved because, to a large extent, one’s basic values reflect the conditions that prevailed during one’s pre-adult years.” (1997:33)

Based on these two hypotheses we can form some clear predictions on where and when we will be able to observe a society with large share of postmaterialists. I will expect postmaterialist values to be most widespread in richer societies than in poor societies, and even more in the richest and most secure strata of a society than in the less educated and poorer strata. I can also expect there to be significant differences between the old and young population in societies which have experienced long term growth in the economy. Hence will the intergenerational differences be smaller in countries with low or no growth in economy and security.

Data

This analysis will use data from two main sources, the World Values Survey(2006), and Angus Maddison’s(2009) data on GDP per capita across countries from 1900 – 2006. As I aim to investigate the effect of level of economic development on personal values, I find it very useful to make use of the World Values Survey. This dataset contains, in addition to responses to wide range of questions on different subjects, personal information about socio-economic status, sex, number of children, geographic information age and a lot more. These provide the researcher with endless possibilities for analyzing change in values over time, between countries, between different socio-economical groups or between men and women.

Independent variables

The British economist Angus Maddison has compiled an extensive dataset which contains data on the level of GDP in the countries of the world from year 1 to 2006. This dataset will provide the data for my two most important independent variables. That is GDP per capita in year of survey and GDP per capita in formative years. The first one is the level of economic development, in GDP per capita, in the country where the respondent live measured at the year the survey took place. The second variable is the level of economic development in the same country at the year in history when the respondent was twenty years old. So in the data for my analysis every individual in the WVS have values for GDP in their country at two different points in their life. Hopefully will this article be able to conclude on which of the two is the better predictor of postmaterialistic values. Dalton(2002) tries to test Inglehart’s hypothesis by aggregating the postmaterialist results for each country in 1990 and comparing them with level of GDP in 1965. He explains that “this is roughly when values were being formed for many of the adults surveyed in 1990”(Dalton 2002:95). This analysis will try to replicate that method, with updated data, for so to introduce the more specific variable for GDP in formative year.

Other important control variables will also be introduced in the analysis. This is done to investigate which other factors can increase the probability of an individual within a society at a given time in history will possess postmaterialistic values. Control variables will be sex, income, and age. Some will be used in several models, and some in fewer of my models.

Dependent variables

The dependent variables of this analysis are provided in the WVS. The WVS-data includes two variables measuring postmaterialism. They are constructed in a similar way, but the 12-item index is somewhat more comprehensive and robust than the 4-item index. The values of these variables are based on the answers the respondents give to questions on what they think should be preferred goals for their country in the next ten year period. Each question has four alternative answers, out of which the respondent must name two preferences. Out of the four alternatives, two are typical material values and two are typical postmaterial values. In the 4-item index, respondents are given one such question battery with a total of 4 answer alternatives, and in the 12-item index, the respondents are given 3 question batteries with a total of 12 alternatives answers. Based on the responses in these questions the respondents are given a score from zero to five in the 12-item index and from one to three in the 4-item index.[1]

Dalton (2002) suggests that the 4-item index may be to narrow to tap a broad dimension of human values. It may also be sensitive to short term inflation rates, as one of the items it deals with is “fight rising prices”. Its advantage is that it has been used in European value surveys for several decades, so it makes it easy to compare changes in values over time. In the binary regression analysis I will make use of a dummy as a dependent variable in the binary regression. I recode the 4- item postmaterialism index and recode it so that all those categorized as postmaterialists get the score 1, and the rest 0. a similar variable is made with materialist being 1 and the rest 0. These two dummy variables are further used to create the PM variable in the scatter diagram. I create that one by subtracting percentage materialists from postmaterialists in each country/year.

Methods

This analysis will take advantage of different forms of tools to explore the effects of Economic development on postmaterialism. Initially I will make use of an ordinary least squares (OLS) regression to explore the effect of my new variable measuring level of economic development in formative year and compare it to the effect of the regular economic development variable.

Further on I will make use of more sophisticated regression techniques, such as fixed-effect models, to find out if the findings I make in the first round are robust, and a binary logistic regression model to calculate the probability that a person will be in the postmaterialist category given a set of fixed variables.

Analysis

In this section of the article will I present and discuss the statistical findings I have done when investigating the data material. First I will try to replicate some of Ingleharts main findings. It seems like a very difficult task to create a prefect replication, because of the availability of data, but I have strived to do an analysis as similar as possible. I will go on to discuss some preliminary regression results that will guide the way for more robust and sophisticated regression analysis.

[pic]

Figure 2 Scatter dot diagram, Postmaterialism and GDP

”pm” is a measure of the 4-item postmaterialism index. It is constructed by subtracting percentage of materialist from percentage of postmaterialists. Gdp_pc_form_year_mean is a measure of economic development. In this case it is an aggregated number for each country year of GDP for year of the WVS. This is a close replication of the analysis by Inglehart (1997:151)

Inglehart use aggregated data for countries for many of his analyses. In the above figure I have tried to replicate one of his models. Inglehart use one observation for each country, while I use several observations. Ingleharts naturally use only one observation of GDP for each country, and I have used one observation for each country-year. Consequently, I have a higher number of observations in my scatter diagram, but the trend is very much the same as in Ingleharts diagram. The relationship between postmaterialism and economic development is a clear and quite linear. An important assessment to consider is whether this linear relationship is due to changes over time within each country, or due to changes between countries. As I will show in a later, descriptive analysis, there is no clear development towards more postmaterialism in society in general. In the period investigated, from the early 1980s to the mid 2000s, the level of postmaterialism in the investigated countries first declined, before an increase in the last decade. It is then a fair assessment to suggest that all the linear relationship in the diagram above is mostly due to differences between countries, because all countries studied have had an increase in GDP in the period studied, and, as I show in figure 4, the share of postmaterialists have not increased significantly. These considerations may be somewhat conflicting with the theories of Inglehart, but in the fixed effects regression analysis’ (figure 7) I will control for the effects of both time and countries, hence, I can conclude with greater confidence on this matter.

[pic]

Figure 3 Correlation matrix

In a regression analysis it is important to check for multicollinearity. That implies a linear correlation between two or more of the explanatory variables. Large multicollinearities weaken the explanatory power of the model because the variables explain the same variation of the dependent variable. In correlation matrix below are all the correlations between the variables I will use in the regression analysis. We find high correlations between education level and both income (0.33) and socio economic status (0,45), these are to be expected and hence, these variables will not be used at the same time in the analysis. The two post materialist variables correlate, and so do the two GDP variables, this is unproblematic as they are supposed to measure almost the same. Other interesting findings from the correlation matrix include a strong and significant correlation between the GDP-variables and the postmaterialist variables. GDP in formative year seem to have the strongest correlation with GDP and it has a correlation coefficient of 0,27 with the 12 item postmaterialist index. These findings are encouraging because they support my initial theory.

Figure 4 shows some descriptive statistics for the two dependent variables that will be used in this article. It is divided into different “waves” in which the world values survey has taken place. We find that there it is impossible to observe any increase in share of postmaterialists in the surveys. In fact there even appears to be a slight decline in postmaterialists over the years.

|  |  |Postmaterialist 4 item index (range 1-3) |

|  |N Valid |Mean |Std dev |Percent in PM category |

|Wave 1, 1981 - 1989 |22093 |1,81 |0,65 |13,5 |

|Wave 2, 1990 - 1994 |57635 |1,87 |0,65 |15,1 |

|Wave 3, 1995 - 1998 |71320 |1,76 |0,64 |11,3 |

|Wave 4, 1999 - 2005 |57596 |1,76 |0,61 |9,6 |

|total |245961 |1,8 |0,64 |12,4 |

|  |  |Postmaterialist 12 item index (range 0-5) |

|  |N Valid |Mean |Std dev |Percent in PM category |

|Wave 1, 1981 - 1989 |2084 |2,58 |1,28 |8,3 |

|Wave 2, 1990 - 1994 |54173 |2,24 |1,23 |3,8 |

|Wave 3, 1995 - 1998 |64892 |1,9 |1,22 |2,1 |

|Wave 4, 1999 - 2005 |54175 |1,92 |1,18 |1,7 |

|total |175324 |2,02 |1,22 |2,6 |

Figure 4 Descriptive statistics

It must be noted that the 12 item index was only in use in a few countries in the first wave. Therefore will it be difficult to conclude that these findings are valid for the entire population.

As previously mentioned, the findings done in this figure 4 hints to an assumption that it might be the variation between countries, rather than variation over time that provides the development we find in the scatter diagram in figure 2.

In the following regression analysis will the relationship between postmaterialism and economic development be investigated more closely, and I will introduce the variable that separates my research from Ingleharts’. That is the variable measuring the level of economic development for each individual in the WVS in their country at the time when they were 20 years old. This will let me know whether we can detect a more accurate explanation of postmaterial values.

The main findings I do in the preliminary analysis (Figure 5) is a confirmation that there is a positive relationship between level of economic development in a society and postmaterialistic values. Both GDP-variables have a positive effect on the level of postmaterialism. In three out of four regression models where both variables are entered, I find that the effect of GDP in formative years is relatively a lot stronger than the original GDP-variable. This supports Ingleharts socialization hypothesis, which states that values are crystallized early in life, and the conditions in a society at that point in life contributes to form people’s values. Further, I find that also the level of personal income, this shows that the scarcity hypothesis can applied not only to societies, but also to individuals. This is very much in line with the theories of Maslow and his hierarchy of needs. People will need to get their basic physiological and safety needs covered before focusing on values represented by the postmaterialism index. However, I also observe that the R² for all of these models are relatively small (all models have R² below 0.1). This indicates that there is a great deal of variation in the postmaterialism-variables that cannot be explained by the explanatory variables included in the regression models.

|  |Model 1 |Model 2 |Model 3 |Model 4 |Model 5 |

|Dependent variable: |12 Item |4-item |12 Item |12 Item |12 Item |

|  |  | | | |  |

|Model Variables: |  | | | |  |

|  |  | | | |  |

|Constant |1.692* |1.610* |1.590* |1.0901* |1.666* |

|  |(.005) |(.003) |(.008) |(.019) |  |

|GDP survey year1000$ |.016* |.008* |.016* |.034* |-- |

|  |(.001) |(.000) |(.001) |(.001) |  |

|GDP formative1000 $ |.046* |.024* |.044* |.019* |.062* |

|  |(.001) |(.000) |(.001) |(.002) |(.001) |

|Sex |--- |--- |--- |-.018* |-.013 |

|  |  | | |(.007) |(.007) |

|Age |--- |--- |--- |-.007* |.000* |

|  |  | | |.000 |(.000) |

|Scale of income |--- |--- |.026* |.024* |.027* |

|  |  | |(.001) |(.001) |(.001) |

|  |  | | | |  |

|R²: |.075 |.069 |.076 |.081 |.073 |

Figure 5. Linear regression analysis.

Std. errors. in parenthesis, *= significant at 0.05 –level

I will now investigate the findings we have done in the regression analysis a little bit closer by making use of some more advanced analysis techniques. I will firstly make use of a binary logistic regression model in which the dependent variable is recoded as a dummy variable, and the independent variables will predict the probability that the value of the dependent, postmaterialism-variable will be 1, the other variables hold constant. I will secondly make use of some fixed effects models, to check how robust my findings are. By holding the effects of some of the explanatory variables constant, we only take into account those effects that are independent of the other variables. These two methods will reveal any flaws in the analysis performed so far. And I can with greater confidence conclude on whether my results significantly differ from Ronald Ingleharts findings.

|  |Model 1 |Model 2 |Model 3 |Model 4 |Model 5 |

|Dependent variable: |PM dummy |PM dummy |PM dummy |Mat dummy |Mat dummy |

|  | | | | |  |

|Model Variables: | | | | |  |

|  | | | | |  |

|Constant: |-2.264 |-1.667 |-2.916 |-.099 |-.506 |

|  |(.104) |(.189) |(.054) |(.905) |(.603) |

|  | | | | |  |

|GDP formative year 1000$ |.093 |--- |.023 |-.041 |-.118 |

|  |(1.097) | |(1.024) |(.960) |(.889) |

|  | | | | | |

| |---- |.079 |.050 |-.059 | --- |

|GDP survey year 1000$ | | | | | |

|  | |(1.082) |(1.052) |(.943) |  |

|  | | | | |  |

| |-.164 |-.195 |-.093 |.142 |.212 |

|Sex | | | | | |

|  |(.849) |(.823) |(.911) |(1.153) |(1.237) |

|  | | | | |  |

| |.000 |-.020 |-.006 |.005 |.000 |

|Age | | | | | |

|  |(1.000) |(.980) |(.995) |(1.006) |(1.000) |

|  | | | | |  |

| |--- |--- |.013 |-.018 | --- |

|Scales of income | | | | | |

|  | | |(1.013) |(.982) |  |

|  | | | | |  |

| |---- |--- |.135 |-.097 | --- |

|Education level | | | | | |

|  | | |(1.145) |(.907) |  |

|  | | | | |  |

|Nagelkerke R²: |.057 |.072 |.080 |.104 |.079 |

Figure 6. Binary logistic regression

Exp (B) in parenthesis, and Standard errors in brackets The dependent variable PM dummy is constructed by assigning the score 1 to every individual with the high score 3 in the 4 item postmaterialism index, all other individuals are assigned 0. The dependent variable mat dummy assigns 1 to all individuals who score the low score 1 on the 4 item postmaterialism index and 0 to all other individuals.

The interpretations of these results are a little bit different than in the linear regression analysis. The numbers in parenthesis are not standard deviation here, but expected B. This means that they indicate the increase or decrease in odds that a person will be in the category 1 of the dependent variable given an increase in 1 unit of the independent variable. An exp (B) is larger than 1, indicates an increase in the odds, while a value lower than 1 indicates a decrease in the odds that the value of the dependent variable will be 1, the other variables hold constant. In model 1 in the figure above “GDP in formative year “have an Exp (B) of 1.097. Hence, we can conclude that the odds of a person being a postmaterialist increases by 9,7 % for a 1000$ increase in GDP in his or her formative year, if all other variables are held constant. In this analysis I find some of the same main effects that was observed in the linear regression analysis. An increase in any of the GDP-variables will increase the probability of being a postmaterialist, and decrease the probability of being a materialist. The two variables are approximately equally strong. I can se in model 1 and 2 that GDP in formative year have a somewhat stronger effect than GDP in survey year when they are used in identical models. However, when they are used together in the same model (see model 3 and 4), we find the opposite result, that GDP in survey year seem to provide the strongest effect on postmaterialism. These contradictory results make it difficult to conclude on whether the GDP- variable that I introduced has any significant explanatory power. The rest of the explanatory variables also provide results that are less than surprising, considering the results we found in the preliminary regression analysis.

|  |Model 1 |Model 2 |Model 3 |Model 4 |Model 5 |

|Dependent variable: |12- Item |4-item |4-Item |12-Item |12-Item |

|  |  | | | |  |

|Model Variables: |  | | | |  |

|  |  | | | |  |

|Intercept |1,7937* |1,8808* |2,0092* |2,1422* |1,9021* |

|  |(0,0275 |(0,0207) |(0,0288) |(0,404) |(0,0577 |

|GDP survey year1000$ |0,0092* |0,0266* |0,0005 |0,0077 |0,086* |

|  |(0,0008) |(0,0010) |(0,0022) |(0,0043) |(0,0064 |

|GDP formative1000 $ |0,0482* |0,0044* |0,0044* |0,0114* |0,0151* |

|  |(0,0011) |(0,0007) |(0,0007) |(0,0016) |(0,0017) |

|Sex |--- |-0,0681* |-0,0677* |-0,0353* |-0,0346* |

|  |  |(0,0029) |(0,0029) |(0,0064) |(0,0064 |

|Age |--- |-0,0048* |-0,0048* |-0,0097* |-0,0090* |

|  |  |(-0,0002) |(0,0001) |(-0,0003) |(0,0003) |

|  |  | | | |  |

|Country dummies |No |Yes |Yes |Yes |Yes |

|Year dummies |Yes |No |Yes |No |Yes |

|  |  |  |  |  |  |

Figure 7 Fixed effects models

Std. errors in parenthesis, * = significant at .005 level

According to model 2, we find that he odds of being a postmaterialist is about 18% lower for females than for males, and the odds decreases with about two percent for every year of age of the respondent. I can hereby confirm that both education level and income level increase the odds of being in the postmaterialist category.

After completing the binary logistic regression I will move on to the fixed effects model where I will put my findings through a tough test to find out whether my results are robust and credible. When adding dummy variables for both years and countries we control for any differences that occur between countries or any change over time, which means that only the variation within countries at the different points in time is accounted for. This method is one of the strictest tests which you can apply to regression analyses.

What I find when we introduce country and year dummies in the fixed effects models (figure 7) is that in general most effects are weaker than in the linear regression analysis. This is because I remove the unwanted variation of the explanatory variables. The directions of the effects are still the same as in the previous analyses, with positive effects of Both GDP variables and negative effects of sex and age. This confirms that postmaterialism is a feature more often prominent for younger people and males. The effect of the two GDP-variables varies from model to model and it seems difficult to give a conclusive interpretation of them. It is worth to note that the “GDP in formative years” –variable is significant at the 5% level in all five models, while the “GDP in survey year”-variable fails to pass the significance test I two of the five models. In my most robust test, when I include dummy variables for every country and for every year in the period (model 4), I find that GDP in formative years have somewhat stronger impact on postmaterialism than GDP in survey year, and the impact of the first GDP variable is also significant, while the impact of the latter is not.

Conclusions

Ronald Ingleharts theories have been put to the test in this article. My aim was to elaborate on Ingleharts analyses by introducing a new explanatory variable measuring the level of economic development for each person in the world value survey in their country at the time of their when their values were being formed. Inglehart discusses this effect in his books, but never puts it to the test in the way I have done in this article. I have made several interesting findings in this article and will try to summarize some of them here.

My analyses shows that the apparent linear relationship between economic development and level of postmaterialism (figure 2), can for a great part be attributed to variation between countries rather than developments over time. This is a plausible inference because the scatter diagram shows that there is a positive relationship between economic development and postmaterialism, but still we observe a relative decline of postmaterialist over the period investigated (figure 4). The correlation matrix in figure 3 confirms that it is a positive relationship between economic development and postmaterialism, and it also shows that the impact of the GDP-variable variable that I added seems to have a stronger correlation with postmaterialism than the original GDP-variable. In the linear regression analysis (figure5) I find this trend once again. When using both GDP-variables in the same regression, it appears that the effect of GDP in formative years is stronger than GDP in the year of the survey. This is very much in line with the theory that personal values will be influenced more by level of economic development at the time when the person grows up than the current level of economic development.

The interpretations are a bit more difficult in the more complicated regression models. When calculating the odds for being a post materialist or a materialist in the binary regression analysis both the two GDP- variables have the expected direction of influence, but it is difficult to conclude on which of them that has the strongest effect. However, I can conclude that men and young people have higher probabilities of being postmaterialists than women and older people. In the fixed effects model, I put my data to the toughest test available. By holding the variation between countries and over time constant, I only check for the variation within a country at any given year. In this test I once again find that both GDP- variables have a positive impact on postmaterialism. These results are significant and so robust that I conclude that level of GDP both in the formative years and in the year of survey definitely have an impact on postmaterialism. It is, however, not possible to determine which of the two that has the greatest explanatory power.

Even if my findings are both robust and significant, the entire models seem to lack some explanatory power, with very low R² and coefficients. This means that there is yet a great deal of variation in postmaterialism to be explained. In order to sufficiently explain this phenomenon, the analyses should include more variables on cultural and historical data. Such data on various historical and cultural features of states and regions combined with individual data on political preferences and type of area of residence (urban vs. rural) could possibly improve the analysis. Another further analysis which could be interesting with regards to postmaterialism would be to see to what degree level of postmaterialism in a country influences the level of females with parliament or introduction of environmental standards (emission levels and compliance of international environmental agreements). Both gender equality and environmental responsibility can be interpreted as postmaterialistic values, and hence we would expect a positive relationship in these cases.

In the beginning of this article (page 4) I made some predictions on what results I would find the analyses. It turns out that most of my predictions were right. I predicted postmaterialism to be more widespread in richer societies than poorer societies, and even more among the richer strata of the population. As I have already noted, GDP definitely have an impact on postmaterialism, and so does income. So we can conclude that both these predictions were correct. Finally I predicted a higher level of postmaterialism among the young population than the old population. As I have noted before, age is negatively related to postmaterialism, so the probability of being a postmaterialist decreases the older a person is.

Bibliography

Dalton, R. J. (2002). Citizen politics: public opinion and political parties in advanced industrial democracies. New York, Chatham House.

Duch, R. M. and M. A. Taylor (1993). "Postmaterialism and the Economic Condition." American Journal of Political Science 37(3): 747-779.

Hellevik, O. and O. Børretzen (1996). Nordmenn og det gode liv: Norsk monitor 1985-1995. Oslo, Universitetsforl.

Inglehart, R. (1977). The silent revolution: changing values and political styles among Western publics. Princeton, NJ, Princeton University Press.

Inglehart, R. (1997). Modernization and postmodernization: cultural, economic, and political change in 43 societies. Princeton, N.J., Princeton University Press.

Inglehart, R. and P. R. Abramson (1999). "Measuring Postmaterialism." The American Political Science Review 93(3): 665-677.

Knutsen, O. (1979). Forandringer i politiske verdier og politisk dyktighet i norske industrisamfunn: en test av Ronald Ingleharts teser om "den stille revolusjonen" i den norske politiske kulturen. R. Inglehart. Oslo, Universitetet.

Knutsen, O. (1989). "The priorities of Materialist and Post-Materialist Values in the Nordic Countries - a Five-Nation Comparison." Scandinavian Political Studies 12(3): 221-241.

Maddison, A. (2009). Statistics on World Population, GDP and Per Capita GDP, 1-2006 AD. A. Maddison, University of Groningen.

World Values Survey, (2006). XWVSEVS_1981_2000_v20060423.sav, World Values Survey.

-----------------------

[1] The question battery for the 4-item index looks like this “There is a lot of discussion about what the aims of this country should be for the next ten years. On this card are listed some of the goals which different people would prefer. Would you please say which one of these you, yourself, consider most important? What is your second choice? And what is your third choice?

1Maintaining law and order in the nation? 2. Giving the people more to say in important government decisions? 3. Fighting rising prices? 4. Protecting freedom of speech?”

................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download