3.4Variables
Roughly 31 % of the games were played on Saturdays, about 27 % on Thursdays and about 25 % on Tuesdays. In addition to that, a few games were played on Mondays (< 3 %), Wednesdays (< 5 %), Fridays (> 6 %) and Sundays (> 3 %). Somewhat more often there were Monday games in bigger towns (r = 0,109) against teams from far away (r = 0,106). There seems to have been more games on Fridays in bigger towns (r = 0,121) and that seems to have been reducing Saturday games (r = -0,128). Otherwise the weekday variables do not seem to correlate with other variables.
Table 3 : Variables, measurement, source and expected sign
variable
|
Measure
|
Source
|
expected sign
|
Game specific factor
|
home town population, monthly(logHPop)
|
Population Register Centre
|
+
|
|
visitor’s town population, monthly (logVPop)
|
Population Register Centre
|
+
|
|
distance between home town and visitor’s town (logDist)
|
Stadium address
http://www.sm-liiga.fi
distance:
http://kartat.eniro.fi
|
-
|
|
home team’s points per game (logHPoin): if zero, then replaced by 0,01
|
own calclulations based on Jääkiekkokirja 2007-2008
|
+
|
|
visitor’s points per game (logVPoin): if zero, then replaced by 0,01
|
own calclulations based on Jääkiekkokirja 2007-2008
|
?
|
|
home team’s points from 3 last games (logHLast): if zero, then replaced by 0,01
|
own calclulations based on Jääkiekkokirja 2007-2008
|
+
|
|
visitor’s points from 3 last games (logVLast): if zero, then replaced by 0,01
|
own calclulations based on Jääkiekkokirja 2007-2008
|
?
|
Incomes
|
Household’s annual incomes, years 2007 (for Fall season) and 2008 (spring season), NUTS4
|
Statistics Finland
|
?
|
Consumer Confidence Index
|
Consumer Confidence index (CCI), monthly, CCIM = CCI for men, CCIW = CCI for women
|
Statistics Finland
|
?
|
Unemployment
|
regional (NUTS4) unemployment rate (Unempl)
|
http://www.tem.fi
|
?
|
Season specific factor
|
played games since the beginning of th season (logHGame) if zero, then replaced by 0,01
|
own calclulations based on Jääkiekkokirja 2007-2008
|
-
|
ticket price, pjk
|
ticket price (logPrice)
|
Jääkiekkokirja 2007 - 2008
|
-
|
Time specific factor, temperature, tempit
|
temperature at nearist observation site (temp)
|
http://www.tutiempo.net/
|
?
|
Temperature difference
|
Daily average temperature in the nearest observation station, years 1960-1990
|
Tilastoja Suomen ilmastosta 1961-1990 - Climatological
statistics in Finland 1961-1990
|
?
|
Time specific factor, weekday
|
weekday, three dummies TU (tuessay) TH thursday), SA (saturday)
|
|
TU –
TH –
SA +
|
Descriptive statistics and correlation on variables (before taking logarithms) are shown in table 3-3.
Table 3 : Variables, means, standard deviations and correlation matrix.
Var
|
mean
|
std
|
ATT
|
Price
|
Dist
|
Temp
|
TempDiff
|
Inc
|
CCI
|
CCIM
|
CCIW
|
Unempl
|
HomeP
|
VisP
|
HomeG
|
HPoints
|
VPoints
|
H3Last
|
V3Last
|
ATT
|
5014
|
1712
|
1
|
0.620
|
-0.074
|
-0.013
|
-0.037
|
0.576
|
0.295
|
-0.067
|
-0.044
|
-0.601
|
0.730
|
0.138
|
0.052
|
0.313
|
0.022
|
0.292
|
0.034
|
Price
|
25.4
|
4.28
|
|
1
|
-0.212
|
0.045
|
0.014
|
0.619
|
0.421
|
-0.028
|
-0.024
|
-0.681
|
0.833
|
0.109
|
0.023
|
0.125
|
-0.001
|
0.200
|
0.063
|
Dist
|
245
|
154
|
|
|
1
|
-0.081
|
0.015
|
-0.123
|
-0.123
|
-0.044
|
-0.049
|
0.204
|
-0.156
|
-0.154
|
0.027
|
0.065
|
0.071
|
0.012
|
0.054
|
Temp
|
4.51
|
5.56
|
|
|
|
1
|
0.166
|
0.087
|
0.581
|
0.697
|
0.709
|
-0.175
|
0.046
|
-0.001
|
-0.714
|
-0.113
|
-0.107
|
-0.089
|
-0.117
|
TempDiff
|
6.31
|
4.15
|
|
|
|
|
1
|
-0.013
|
-0.287
|
-0.358
|
-0.399
|
0.132
|
0.003
|
-0.011
|
0.422
|
0.088
|
0.054
|
0.098
|
0.064
|
Inc
|
35105
|
3334
|
|
|
|
|
|
1
|
0.615
|
-0.005
|
-0.000
|
-0.849
|
0.838
|
-0.016
|
-0.014
|
0.266
|
-0.005
|
0.282
|
0.002
|
CCI
|
14.7
|
3.59
|
|
|
|
|
|
|
1
|
0.698
|
0.707
|
-0.625
|
0534
|
-0.011
|
-0.691
|
0.083
|
-0.066
|
0.098
|
-0.076
|
CCIM
|
15.5
|
2.40
|
|
|
|
|
|
|
|
1
|
0.892
|
-0.100
|
-0.020
|
-0.004
|
-0.904
|
-0.143
|
-0.122
|
-0.125
|
-0.133
|
CCIW
|
13.5
|
2.99
|
|
|
|
|
|
|
|
|
1
|
-0.103
|
-0.021
|
-0.008
|
-0.926
|
-0.092
|
-0.090
|
-0.078
|
-0.086
|
Unempl
|
8.34
|
2.03
|
|
|
|
|
|
|
|
|
|
1
|
-0.788
|
0.014
|
0.117
|
-0.164
|
0.040
|
-0.190
|
0.036
|
HomeP
|
185356
|
168315
|
|
|
|
|
|
|
|
|
|
|
1
|
-0.012
|
0.010
|
0.192
|
-0.010
|
0.228
|
0.012
|
VisP
|
185121
|
168074
|
|
|
|
|
|
|
|
|
|
|
|
1
|
0.003
|
0.009
|
0.171
|
0.026
|
0.223
|
HomeG
|
27.5
|
16.2
|
|
|
|
|
|
|
|
|
|
|
|
|
1
|
0.120
|
0.119
|
0.106
|
0.131
|
HPoints
|
1.46
|
0.511
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1
|
0.224
|
0.644
|
0.086
|
VPoints
|
1.48
|
0.530
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1
|
0.151
|
0.643
|
H3Last
|
4.24
|
2.51
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1
|
0.099
|
V3Last
|
4.42
|
2.52
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1
|
ATT = attendance, Price (€), Dist = distance between home team’s and visitor’s stadiums along road (km), Temp = max tempature, TempDiff = Av. Temp – max temp, Inc = incomes, CCI = consumer confidence index, CCIM = CCI of men, CCIW = CCI of women, Unempl = monthly regional unemployment rate(%), HomeP= home town population, VisP = vistor’s town population, HomeG = number of games, home team, before the game, HPoints = points per game, home team, before the game VPoint = visitor’s points per game, before the game, , VisiG = number of games, visitor, before the game , Last3H = points from 3 last games, home team, Last3V = points from 3 last games, visitor. The number of observations = 392
|
The correlation matrix reveals the ticket price seems to have been higher in larger towns and it seems to have a positive relation with attendance which is positively related to incomes (households’ annual average incomes in the NUTS4-region). Since the regular season begins in September and ends in March and the income variable is annual, although different in the fall season and spring season, the possible bias is corrected with monthly consumer confidence index (CCI). There are three alternatives: CCI and CCI for men (CCIM) and for women (CCIW). CCI for men (CCIM) should have a bigger impact on attendance than the other two alternatives. The number of home team games and the number of visitor’s games were (naturally) highly positively correlated (not shown). Points per game from the beginning of the season (HPoint) and points from the last three games (Last3H) were also positively correlated. The regional unemployment rate seems to have been higher in areas with smaller towns. The temperature seems to have been lower when the number of games has increased. Probably the relation is like inverse U or inverse J. According to long-term statistics (1900 – 2000) the temperature in Helsinki (Kaisaniemi observation site) has been + 11,1 Celsius in September, + 6,2 Celsius in October, + 1,5 Celsius in November, - 2,1 Celsius in December, -4,7 Celsius in January, -5,7 Celsius in February and -2,2 in March (Ilmatieteen laitos 2009). Due to possible bias in interpretation the temperature difference (TempDiff) are measured as devation from these long-run averages.
Since points per game (HPoints or VPoints) and the corresponding points from the last three games (3HLast or V3Last) are strongly positively correlated and these partially measure the same for empirical purposes, these are used as alternative measures.
Share with your friends: |
