// How reasistitic are the evaluations about countries Living Condions version 8 set more off set scheme s1mono capture log close log using anvalid1D, replace // Data // ---- use ID cntry year hungary_i-turkey_i hungary_g-turkey_g using data02, clear label define country 4 "Germany (W)" 3 "Germany (E)" 2 "Hungary" 1 "Turkey" encode cntry, gen(country) label(country) replace germany_g = otherpart_g if cntry == "Germany (W)" | cntry == "Germany (E)" replace germany_i = otherpart_i if cntry == "Germany (W)" | cntry == "Germany (E)" replace hungary_g = . if cntry=="Hungary" replace hungary_i = . if cntry=="Hungary" replace turkey_i = . if cntry=="Turkey" foreach word in hungary poland spain italy germany france sweden netherlands switzerland { local cap = proper("`word'") label var `word'_g "`cap'" label var `word'_i "`cap'" } collapse (mean) /// switzerland_i netherlands_i sweden_i france_i germany_i italy_i spain_i hungary_i poland_i, /// by(country) local i 9 foreach var of varlist switzerland_i-poland_i { ren `var' mean`i--' } reshape long mean, i(country) j(refgroup) label define refgroup /// 1 Polen /// 2 Ungarn /// 3 Spanien /// 4 Italien /// 5 Deutschland /// 6 Frankreich /// 7 Schweden /// 8 Niederlande /// 9 Schweiz label value refgroup refgroup reshape wide mean, i(refgroup) j(country) format mean* %3.1f // To type text table lines separated by & characters for cutting and pasting into a // Microsoft Word table using the menu sequence Table->Convert->Text to Table: listtex refgroup mean1 mean2 mean3 mean4 /// using anvalid1D.txt /// , replace rstyle(tabdelim) /// head(Türkei Ungarn Deutschland (E) Deutschland (W)) log close exit Two sorts of such comparisons can be made. Firstly, we can subtract the score of the reference country from the score for the own country, and secondly we can subtract the score for the reference country from the score for the evaluation of the individual living condition. These two comparisons refer to the two modes of comparison described above (-> Jan?). By doing the respective subtraction we will get a measure, which can vary between -10 and 10. We will get negative values if the respondent evaluates the his living condition or his own country's living condition to be lower than the living conditions in the reference country, and we will get positive values if it is the other way around. The value of 0 indicates that a respondent evaluates the his living conditions or his country's living condition to be the same as in the reference countries to be the same. Figure [anvalid1a.eps] and [anvalid1b.eps] displays box plots (Cleveland 1994: 139--143) of the distributions of variables generated accordingly. Thereby [anvalid1b.eps] displays the results for the individual mode comparison and [anvalid1b.eps] for the general mode comparison (-> check term). Within each figure the distributions are presented in separate panels for each survey country. In each panel of the figure the reference countries are sorted according to their GDP per capita (in PPP). It can be seen from figure [anvalid1a.eps] that the mass of the respondents from Turkey evaluates the their living conditions to be substantially lower than the living conditions in any other country. Even in the comparison to the poorest of the reference countries --- Poland --- about 75 percent of the respondents from Turkey believe that their living condition is worse than that in Poland. Hungarian respondents by and large share the pessimistic views of the Turks, i.e. they belive that their living condition is worse than the living conditions in the Western European reference countries. However, Hungarian respondents evaluate their living conditions to be better than that of an average Pole. Refer to the panels for Germany, special attention is necessary for the comparison of Germans with "Germany". In the survey, both, East-Germans and West-Germans are ask to evaluate the living conditions in western and eastern Germany separately. Therefore we are able to measure how West-Germans evaluate their living condition in comparison to East Germany, and how East Germans evaluate their living conditions in comparison to West Germany. It can be seen from the figure, that the respondents from both parts of Germany consider the living conditions in West Germany to be favorable to those in East Germany. Besides that respondents from both parts of Germany tend to see their living conditions worse than the living conditions in Switzerland, better than the living conditions in Poland, Hungary and the bulk of western European countries and similar than the living conditions in France. A very similar picture arise if we look at the distributions of the variables for the general mode of comparison (figure [anvalid1b.eps]). However there is one exception, which is worth noting: The respondents from East Germany seem to be slightly too patriotic in the evaluation of East German's living conditions. They evaluate the living conditions to be similar than those in Switzerland and yet better than those in the Netherlands. According to their answers, only West-Germany has substantial better living conditions than East Germany. A methodological explanation for these somewhat unrealistic responds might be, that the question on evaluating the living conditions in eastern Germany is answered with respect to the living conditions in the entire country, and that the answers about West-Germany are calibrated accordingly. Besides this minor glitch, two general points that are important for subsequent analysis: - The evaluations of the respondents by and large reflect the position of the reference countries in the GDP league table of nations. In this sense the evaluations can be seen as a realistic picture of an objective entity. However even if the evaluations were unrealistic, this would not invalidate an analysis of the question whether the --- possibly wrong --- evaluations of ``international'' reference groups affect peoples live satisfaction. - The distributions of the comparison-measures vary substantially within each survey country, and most of the distributions seem to be fairly symmetric. Hence, there is no barrier to use these variables as independent variables from the outset.