4A
DATA & SAMPLE
I used the General Social Survey which contains demographic, behavioral, and attitudinal questions, plus topics of special interest (University of Chicago, n d). The survey has remained generally the same since 1972 so that time-trend studies over cohorts and generations could be analyzed. To observe a single appropriate snapshot of current responses, I used the 2006 data set. For the data set, there were 4,510 observations present in the sample. Table 1 shows the number of observations for each of the respective variables.
TABLE 1: . sum tvhours attend
Variable | Obs Mean Std. Dev. Min Max
————————————————————————————————————————————–
tvhours | 1987 2.935581 2.285925 0 24
attend | 4491 3.568693 2.797691 0 8
4B
MEASURES
I studied the relationship between church attendance and watching TV in the United States. Church attendance, the independent variable, was treated as ordinal. In the survey, there were nine total answer options ranging from “never” to “more than one per week” based on how often the subject reported to attend religious services. The data was recoded to differentiate between low and high frequency church attendance. The variable ATTEND was recoded with responses “never to once a year” as low attend church attendance and “several times a year to more than once a week” as high attend church attendance. This distinction was appropriate because church attendance generally rises on Christmas (or Easter) and many non-attenders do go on one of these holidays (Barnett, 1976). Table 2 shows a frequency table of church attendance. The dependent variable, TVHOURS, was treated as a continuous variable because there were twenty-four separate and continuous survey response options that follow the number of hours in a day from one to twenty-four. The dependent variable measured number of hours the respondent spent generally watching TV per day.
TABLE 2: . tab attendR
Church |
Attendance | Freq. Percent Cum.
————————————————————————————————————————————–
Low Attend | 1,893 42.15 42.15
High Attend | 2,598 57.85 100.00
————————————————————————————————————————————–
Total | 4,491 100.00
4C
METHOD
Regression, chi-square, and gamma tests will be used to show whether there is a relationship between church attendance and TV watching in the United States. Table 3 shows regression.
TABLE 3: . reg tvhours attendR
Source | SS df MS Number of obs = 1979
—————————————————————————————-
. F( 1, 1977) = 9.31
Model | 48.1273161 1 48.1273161 Prob > F = 0.0023
Residual | 10221.5938 1977 5.17025481 R-squared = 0.0047
—————————————————————————————-
. Adj R-squared = 0.0042
Total | 10269.7211 1978 5.19197223 Root MSE = 2.2738
——————————————————————————————————————————
tvhours | Coef. Std. Err. t P>|t| [95% Conf. Interval]
——————————————————————————————————————————
attendR | -.3152709 .1033342 -3.05 0.002 -.5179262 -.1126155
_cons | 3.115976 .0782218 39.84 0.000 2.962571 3.269382
——————————————————————————————————————————
Regression
Table 3 shows OLS regression data that can be used for analyzing the estimated effects of church attendance on hours of TV watching per day. It shows that when church attendance is 0 (constant), hours of watching TV per day is 3.11. As church attendance raises by 1 unit, I predict that hours of TV watching per day will decrease by .315 units (negative slope). Because the slope is -0.3.15 (negative) it indicates a negative association. This is significant at alpha = 0.05 level. The R-squared value of 0.0047, variance in hours watching TV per day, can be explained by differences in church attendance. When church attendance is low (x=0) we predict TV watching hours to equal 3.11. When church attendance is high (x=1) we predict TV watching hours to equal 2.795 hours (Y=3.11+(-.315)X). Figure 1 provides a visual representation of the data
Figure 1: . plot tvhours attendR
24 +
h | *
o |
u |
r |
s |
- | *
p |
e |
r | * *
- | *
d | * *
a |
y | * *
- | * *
w | * *
a | * *
t | * *
c | * *
h | * *
0 + * *
——————————————————————————————————————————
0 Church Attendance 1
Chi-square Test
I performed a chi-square test (Table 4) to show whether or not there was a relationship between TV watching and church attendance. With a P-value at .049, I reject H0 at the alpha ? = 0.05 level because .045<.05 so there is no relationship at the .05 alpha level. The x2 (chi-square) with df = 16 (degree of freedom) is x2=26.4 which indicates a relationship between TV watching and church attendance. The gamma value is -0.09 which indicates a slight negative relationship.
Table 4: . tab tvhours attendR, col chi gamma
——————————————————————————————————————————
| Key |
——————————————————————————————————————————
| frequency |
| column percentage |
——————————————————————————————————————————
hours per |
day |
watching | Church Attendance
tv | Low Atten High Atte | Total
——————————————————————————————————————————
0 | 30 48 | 78
. | 3.55 4.23 | 3.94
——————————————————————————————————————————
1 | 162 257 | 419
. | 19.17 22.66 | 21.17
——————————————————————————————————————————
2 | 244 331 | 575
. | 28.88 29.19 | 29.06
——————————————————————————————————————————
3 | 132 205 | 337
. | 15.62 18.08 | 17.03
——————————————————————————————————————————
4 | 108 118 | 226
. | 12.78 10.41 | 11.42
——————————————————————————————————————————
5 | 72 63 | 135
. | 8.52 5.56 | 6.82
——————————————————————————————————————————
6 | 39 60 | 99
. | 4.62 5.29 | 5.00
——————————————————————————————————————————
7 | 12 11 | 23
. | 1.42 0.97 | 1.16
——————————————————————————————————————————
8 | 18 16 | 34
. | 2.13 1.41 | 1.72
——————————————————————————————————————————
9 | 2 2 | 4
. | 0.24 0.18 | 0.20
——————————————————————————————————————————
10 | 12 11 | 23
. | 1.42 0.97 | 1.16
——————————————————————————————————————————
12 | 6 7 | 13
. | 0.71 0.62 | 0.66
——————————————————————————————————————————
13 | 0 1 | 1
. | 0.00 0.09 | 0.05
——————————————————————————————————————————
14 | 6 1 | 7
. | 0.71 0.09 | 0.35
——————————————————————————————————————————
15 | 1 1 | 2
. | 0.12 0.09 | 0.10
——————————————————————————————————————————
18 | 0 2 | 2
. | 0.00 0.18 | 0.10
——————————————————————————————————————————
24 | 1 0 | 1
. | 0.12 0.00 | 0.05
——————————————————————————————————————————
Total | 845 1,134 | 1,979
| 100.00 100.00 | 100.00
Pearson chi2(16) = 26.3786 Pr = 0.049
gamma = -0.0923 ASE = 0.031
4e
Conclusion
Based on the results using the GSS 2006 data, church attendance may play a significant role on the hours of TV watching per day (individually) in the United States. However, it has not been shown that there is a strong relationship. The relationship is slightly negative. Church attendance explains 0.0047 of the variance in hours of TV watching per day (individually) in the United States.
