Social Research Skills I
 a) Variables may take different levels of measurement including scale (ratio, interval) variables, nominal and ordinal variables. Scale variables are categorized such that it is possible to place variables in ranks of greater to lesser. In addition, such variables have to be numeric, whereas the data values are computable. There is mutual exclusivity of data categories. Examples of scale variables include height and age among others. Ordinal level data is composed of data values that can be arranged in a certain intrinsic order such that it is impossible to define the measure of difference between the categories. Data categories exhibit mutual exclusivity. Ordinal values may include strongly disagree, agree, strongly agree. Finally, nominal data values can not be categorized in any intrinsic order. It is impossible to describe data from greater to lesser. A respondent’s name is the best example of a nominal data value (Healey & Prus, 2010).
 b) The 2008 British Social Attitudes Survey dataset contains ratio, nominal and ordinal levels of measurement. Variables that have ratio measurement include: Serial, age, NumCh, NCh415, NCh318, HhCh04, HhCh511, HhCh1215, HhCH1617, leftrigh, libauth, Age. Nominal: Rsex, WhPaper, GOR2, Country, Religion, ReligSum, Vote05sc, Welfare2, PrSch100, HiPay100, ChPov100, RNSEGGrp, FTREL. Ordinal: IncomGap, MumPoor, Pen1Poor, MumOn130, LpWkAge, WkJbHrsl, UnionSA, LIFESAT, Tea2, FathWk.
 c) Scale level variables help in computing quantitative statistical analyses whereas ordinal and nominal variables help in computing qualitative statistical analyses.
The total number of respondents in the 2008 British Social Attitudes Survey is 4,486 respondents (Table 1). According to Table 1, the mean age of the sample was 50.33 years with a standard deviation of 17.936. Table 2 and Figure 1 show that 57.1 percent of the respondents in this survey were females.
 a) According to Table 1, the youngest respondent is 17 years whereas the oldest respondent is 95 years.
 b) Recoded ages were categorized as follows: 1 represents 1824, 2 for 2534, 3 for 3544, 4 for 4554, 5 for5564 and 6 for 65+.
 c) For respondents who are in the 3544 category, 22.6 percent of the respondents are members of a trade union or a staff association (Table 4).
 a) Table 5 indicates that the minimum hours a respondent works per week is 0 hours whereas the maximum number of hours worked is 90 hours. The mean number of hours worked in a week is 20.47 hours with a standard deviation of 21.135 while the median number of hours a respondent’s works per week is 16.00 hours.
 b) The confidence interval of the mean number of hours respondents work per week is 1.96 to +1.96 since the tests are conducted at 95% level of significance.
 c) Confidence interval (CI) is the range of probability values that a given experiment is said to be valid. Thus, CI helps in determining the validity of a given outcome in an experiment (Leech, Barrett & Morgan, 2005).
 a) To test significance, it is first important to state the null and alternate hypotheses. Secondly, the alpha values are stated, i.e. 1.96 to 1.96 followed by computing the probability values for a given data set. The p values are then compared with the alpha value thus forming a basis for accepting or rejecting the null hypothesis.
 b) When interpreting a statistically significant finding, social scientists should consider the degrees of freedom (df) and the alpha value as well as the probability value (Field, 2009).
The valid number of voters who did not vote in England was 257, those who voted for Conservative in England were 196, those who voted for Labour were 242, LiberalDemocrat voters were 91 while those who voted for other party were 20. In Scotland, 21 voters did not vote, 15 voted for Conservative, 26 voted for Labour, 9 voted for Liberal Democrat while 18 voted for other party. In Wales, 10 voters did not vote, 12 voted for Conservative, 17 voted for Labour, 5 voted for Liberal Democrat while 6 voted for other party (Table 6).
There is no statistically significant association between the respondent’s country of resident and the party they voted for in 2005 general elections. This is because the ChiSquare test indicates that 2 cells had expected count less than 5 (Table 7). This is despite the Pearson ChiSquare 66.348, df 8 and p =.001 being significant (Weinberg & Abramowitz, 2008).
 a). The MannWhitney test indicates that more females are concerned about how old a child should be before the parent goes back to work (Female mean rank of 1354.32 against males mean rank of 1594.67). The difference is statistically significant (MannWhitney U= 865157.00, Z = 7.789 and 2tailed significance of .001) (Table 8).
 According to MannWhitney’s test, females believe that a greater proportion of under 16 are living under poverty i.e. 1820 against 1402 males (Table 8). From Table 9 the difference is statistically significant (MannWhitney U = 1001165.50, Z = 10.539 and 2tailed significance of .001).
The respondent’s mean authoritarian/libertarian scale differs according to how he/she voted. This is indicated by the nonsignificant Levene statistic 1.997, df 4 p =.093. According to Table 11, the ANOVA test showed a significant difference (F = 9.790, df 4, p <.001) on how voters voted as a factor of mean authoritarian/libertarian scale.
 A 2tailed Pearson Correlation between a respondent’s leftright score and the proportion of children under sixteen they believe are living in poverty shows that a statistically significant relationship exists. There is a negative relationship (r_{p}=.244, p =.001, which is significant) between leftright scale ABCD and children 16, who live in poverty (Table 13).
Bibliography
Field, A. P. (2009). Discovering statistics using SPSS. Sixth edition. London: Sage Publications Ltd.
Healey, J. F. and Prus, S. G. (2010). Statistics: A tool for social research. First Canadian edition. Canada: Nelson Education Ltd.
Leech, N. L., Barrett, K. C. and Morgan, G. A. (2005). SPSS for intermediate statistics: use and interpretation. Second edition. New York, NY: Routledge.
Weinberg, S. L. and Abramowitz, S. K. (2008). Statistics using SPSS: an integrative approach. 2^{nd} edition. Cambridge, UK: Cambridge University Press.
Appendix
Table 1: Descriptive Statistics
Statistics  
SEX OF respondent :Q45  What was respondent’s age last birthday? :Q46  
N  Valid  4486  4466 
Missing  0  20  
Mean  1.57  50.33  
Median  2.00  49.00  
Mode  2  42  
Std. Deviation  .495  17.936  
Variance  .245  321.696  
Minimum  1  17  
Maximum  2  95 
Table 2: Respondent’s Sex
SEX OF respondent :Q45  
Frequency  Percent  Valid Percent  Cumulative Percent  
Valid  Male  1926  42.9  42.9  42.9 
Female  2560  57.1  57.1  100.0  
Total  4486  100.0  100.0 
Table 3: Descriptive for Recoded Age
What was respondent’s age last birthday? :Q46  
Frequency  Percent  Valid Percent  Cumulative Percent  
Valid  1  313  7.0  7.0  7.0 
2  661  14.7  14.7  21.7  
3  886  19.8  19.8  41.5  
4  772  17.2  17.2  58.7  
5  749  16.7  16.7  75.4  
6  1103  24.6  24.6  100.0  
17  2  .0  .0  100.0  
Total  4486  100.0  100.0 
Table 4: Cross Tabulation of Respondent’s Age and Membership of a Trade Union
What was respondent’s age last birthday? :Q46 * Are you now a member of a trade union or staff association? Q993 Crosstabulation  
Are you now a member of a trade union or staff association? Q993  Total  
Yes, trade union  Yes, staff association  No  
What was respondent’s age last birthday? :Q46  1  Count  17  7  286  310 
Expected Count  53.5  8.6  247.9  310.0  
% within What was respondent’s age last birthday? :Q46  5.5%  2.3%  92.3%  100.0%  
% within Are you now a member of a trade union or staff association? Q993  2.2%  5.6%  8.0%  6.9%  
% of Total  .4%  .2%  6.4%  6.9%  
2  Count  107  22  529  658  
Expected Count  113.6  18.3  526.2  658.0  
% within What was respondent’s age last birthday? :Q46  16.3%  3.3%  80.4%  100.0%  
% within Are you now a member of a trade union or staff association? Q993  13.9%  17.7%  14.8%  14.7%  
% of Total  2.4%  .5%  11.8%  14.7%  
3  Count  174  31  678  883  
Expected Count  152.4  24.5  706.1  883.0  
% within What was respondent’s age last birthday? :Q46  19.7%  3.5%  76.8%  100.0%  
% within Are you now a member of a trade union or staff association? Q993  22.6%  25.0%  19.0%  19.8%  
% of Total  3.9%  .7%  15.2%  19.8%  
4  Count  199  24  549  772  
Expected Count  133.2  21.4  617.3  772.0  
% within What was respondent’s age last birthday? :Q46  25.8%  3.1%  71.1%  100.0%  
% within Are you now a member of a trade union or staff association? Q993  25.8%  19.4%  15.4%  17.3%  
% of Total  4.5%  .5%  12.3%  17.3%  
5  Count  154  19  575  748  
Expected Count  129.1  20.8  598.1  748.0  
% within What was respondent’s age last birthday? :Q46  20.6%  2.5%  76.9%  100.0%  
% within Are you now a member of a trade union or staff association? Q993  20.0%  15.3%  16.1%  16.7%  
% of Total  3.4%  .4%  12.9%  16.7%  
6  Count  120  21  953  1094  
Expected Count  188.8  30.4  874.8  1094.0  
% within What was respondent’s age last birthday? :Q46  11.0%  1.9%  87.1%  100.0%  
% within Are you now a member of a trade union or staff association? Q993  15.6%  16.9%  26.7%  24.5%  
% of Total  2.7%  .5%  21.3%  24.5%  
17  Count  0  0  2  2  
Expected Count  .3  .1  1.6  2.0  
% within What was respondent’s age last birthday? :Q46  .0%  .0%  100.0%  100.0%  
% within Are you now a member of a trade union or staff association? Q993  .0%  .0%  .1%  .0%  
% of Total  .0%  .0%  .0%  .0%  
Total  Count  771  124  3572  4467  
Expected Count  771.0  124.0  3572.0  4467.0  
% within What was respondent’s age last birthday? :Q46  17.3%  2.8%  80.0%  100.0%  
% within Are you now a member of a trade union or staff association? Q993  100.0%  100.0%  100.0%  100.0%  
% of Total  17.3%  2.8%  80.0%  100.0% 
Table 5: Number of Hours worked per week, including overtime
Statistics  
How many hours R normally work a week including overtime :Q968  
N  Valid  4368 
Missing  118  
Mean  20.47  
Std. Error of Mean  .320  
Median  16.00  
Mode  0  
Std. Deviation  21.135  
Variance  446.669  
Range  90  
Minimum  0  
Maximum  90 
Table 6: Cross Tabulation of Country of respondent and Party voted in 2005 General Elections
Country of interview England, Scotland or Wales? :Q31 * Which party voted for in 2005 general election? D2.25 Crosstabulation  
Which party voted for in 2005 general election? D2.25  Total  
Did not vote  Voted Conservative  Voted Labour  Voted Liberal Democrat  Voted other party  
Country of interview England, Scotland or Wales? :Q31  England  Count  257  196  242  91  20  806 
Expected Count  245.6  190.2  243.1  89.6  37.5  806.0  
Scotland  Count  21  15  26  9  18  89  
Expected Count  27.1  21.0  26.8  9.9  4.1  89.0  
Wales  Count  10  12  17  5  6  50  
Expected Count  15.2  11.8  15.1  5.6  2.3  50.0  
Total  Count  288  223  285  105  44  945  
Expected Count  288.0  223.0  285.0  105.0  44.0  945.0 
Table 7: ChiSquare Test for Country and Party Voted
ChiSquare Tests  
Value  df  Asymp. Sig. (2sided)  
Pearson ChiSquare  66.348^{a}  8  .000 
Likelihood Ratio  45.646  8  .000 
LinearbyLinear Association  18.599  1  .000 
N of Valid Cases  945  

Table 8: MannWhitney Test for Sex and How Old a Child Should Be Before the Parent Goes Back to Work
Ranks  
SEX OF respondent :Q45  N  Mean Rank  Sum of Ranks  
How old should child be before parent work? [shld seek wk] :BCD383  Male  1238  1594.67  1974196.00 
Female  1674  1354.32  2267132.00  
Total  2912 
Table 9: MannWhitney U
Test Statistics^{a}  
How old should child be before parent work? [shld seek wk] :BCD383  
MannWhitney U  865157.000 
Wilcoxon W  2267132.000 
Z  7.897 
Asymp. Sig. (2tailed)  .000 

Table 10: MannWhitney Test for Children under 16 Who May Live In Poverty and Sex
Ranks  
SEX OF respondent :Q45  N  Mean Rank  Sum of Ranks  
Of every 100 children under 16, how many live in poverty? :ACD557  Male  1402  1415.60  1984668.50 
Female  1820  1762.41  3207584.50  
Total  3222 
Test Statistics^{a}  
Of every 100 children under 16, how many live in poverty? :ACD557  
MannWhitney U  1001165.500 
Wilcoxon W  1984668.500 
Z  10.539 
Asymp. Sig. (2tailed)  .000 

Table 11: Levene’s Statistic
Test of Homogeneity of Variances  
Libertarianauthoritarian scale ABCD  
Levene Statistic  df1  df2  Sig. 
1.997  4  932  .093 
Table 12: ANOVA Test for LibertarianAuthoritarian Scale
ANOVA  
Libertarianauthoritarian scale ABCD  
Sum of Squares  df  Mean Square  F  Sig.  
Between Groups  14.524  4  3.631  9.790  .000 
Within Groups  345.658  932  .371  
Total  360.182  936 
Table 13: Pearson Correlation between a Respondent’s LeftRight Score and the Proportion of Children under Sixteen They Believe are Living in Poverty
Correlations  
Leftright scale ABCD  Of every 100 children under 16, how many live in poverty? :ACD557  
Leftright scale ABCD  Pearson Correlation  1  .244^{**} 
Sig. (2tailed)  .000  
N  3902  2830  
Of every 100 children under 16, how many live in poverty? :ACD557  Pearson Correlation  .244^{**}  1 
Sig. (2tailed)  .000  
N  2830  3222  
**. Correlation is significant at the 0.01 level (2tailed).

Figure 1: Percentage of participants according to sex
Figure 2: A bar graph of Recoded Age Categories
Use the following coupon code :
theRUSH