INTRODUCTION
In 1994, at the international meeting of Health Promotion and Physical Activity
held by WHO and FIMS, the WHO presented the concept of sports lifestyle. In
1995, Japanese scholars once again proposed that the sports in the twentyfirst
century are a kind of cultural lifestyle, namely the sports lifestyle. Since,
the Nationwide Fitness Program Outline was put into enforcement in 1995, the
mass sports activities in China have entered a new stage of development and
many scholars discussed the sports lifestyle, as well as the relation between
lifestyle and sports from different angles (Arrow, 1963).
The sports lifestyle is subject to the lifestyle system, being the partwhole
relationship. It is the sum of sports activities and behavior characteristics
for people to meet the requirements of selfmaintenance and development by available
sport resources according to their subjective desire under certain natural conditions
and social background (LeyvaLopez and FernandezGonzalez,
2003). The sports lifestyle focuses on maintaining the exercise habit. The
sports participative behavior is the most important manifestation of sports
lifestyle which is also the key point to measure the sports lifestyles of different
populations at various times.
MATERIALS AND METHODS
The research object is the college students. Currently, the primary task of
sports work in Chinese universities is to improve students’ health, the
purpose of health education is to cultivate good health behavior and exercise
habit for students, so that they can develop physically and mentally to enhance
the social adaptation ability (Hwang and Masud, 1979).
However, the PE once a week in universities can not satisfy the students to
build up body and the problem that the physiques of college students in our
country are declining can not be solved only by PEs. There are various factors
influencing their physical health while the extracurricular sports lifestyle
is the most active and positive factor. The research on sports lifestyle evaluation
indexes is to evaluate from physiology, psychology and sociology (Ngo
and Mousseau, 2002). But since the college students have been during the
course of physical education from primary school to university and their sports
activities and physical training are of some regularity. Different from general
population, their sports lifestyle has its own particularity. This study mainly
studies the extracurricular sports lifestyle of college students.
The research methods of this study can be summarized as follows:
• 
Literature method, consulting a large number of books and
studys related to sports lifestyle both at home and abroad. Understand the
concept, main features and contents of sports lifestyle. And list out the
main contents of extracurricular sports lifestyle on college students,
including the purpose, site facility, project and effect of sports activities
and up to basic standards of sport population, namely the time, frequency
and intensity, respectively reach 30 min at a time, 3 times per week and
above moderate intensity 
• 
The method of interview, consulting experts about the questionnaire contents
and design plan, as well as the contents of extracurricular sports lifestyle
on college students 
• 
Questionnaire survey, developing the questionnaire according to above
contents in five aspects, distributing 100 questionnaires and receiving
95 valid ones, with the valid return rate as 95% 
• 
Mathematical statistics, applying the SPSS 10.0 software to data reduction,
statistics and comparison 
PERFORMANCE INDEX SYSTEM
After the evaluation objective is set, based on theoretical analysis, present
the evaluation indexes determined initially. Design the questionnaire, whose
contents are the purpose, site facility, project and effect of sports activities
and up to basic standards of sport population, namely the time, frequency and
intensity, respectively reach 30 min at a time, 3 times per week and above moderate
intensity. It has a total of more than 30 questions. The reliability and validity
test: The reliability coefficient R of this questionnaire is 0.87 and by expert
judgment, its validity test shows that 80% experts agree with it. It can thus
be seen that, this questionnaire has a good reliability and validity, meeting
the requirements of statistics and research.
Consult the experts by Delphi method and request them to assign all the listed
indexes by the fivegrade scoring method. After the experts complete scoring,
calculate the mean number to obtain the average score for each index and finally
screen out the indexes with the average score ≥4. Count and optimize them
(the second screening). Through experts’
evaluation and based on research need, the indexes were analyzed, respectively
during the second screening. Then they were consolidated and concluded. The
firstgrade indexes are: The objectives of sports activities (fitness, entertainment,
social intercourse); the sports activities process (site facility, sports event,
sports time, frequency and intensity); the effects of sports activities (the
results of fitness, education and social adjustment).
The retest method was employed to test their reliability. Firstly, test them
two times and calculate the interclass correlation coefficients of their test
results to obtain the testretest reliability coefficients. The reliability
coefficients between 0.650.75 are acceptable, fairly good between 0.750.85
and very good above 0.85, representing the degree of reliability for the tested
indexes. The test results show that, the reliability coefficients of all indexes
are within the acceptable range. The test project indexes are highly correlated
to the sports lifestyle on college students. It indicates that the test indexes
are highly associated with the extracurricular sports lifestyle of college
students. Therefore, this study will further analyze these indexes.
In an index system, the importance degrees of various indexes are different.
In order to reflect their degrees of importance when making an evaluation conclusion,
their corresponding weights should be determined. The weights of evaluation
indexes in all grades can be determined by AHP method. (1) The experts sorted
the evaluation indexes by their importance degrees in descending order through
pairwise comparison with their experiences and assigned relatively important
rank values, (2) Construct matrixes to judge various index weight coefficients
and (3) Calculate the weights q of all indexes to obtain the weight coefficients
of the first and second grade indexes, as shown in Table 1.
By testing the index weights, the reliability coefficient is 93% which is satisfactory.
PROPOSED APPROACH
In group decision, there are two ways to evaluate the alternatives, namely
the consistency criterion and the respective evaluation. When the members in
group decision committee can reach consensus on evaluation criteria and their
weights, it can use the consistency criterion. If not, the respective evaluation
should be used. Firstly each member should give the overall merits of alternatives
and then collect their evaluations to form the group decision (Roy
et al., 1986). Since, in the consistency criterion, the ELECTREII
is used just once and the group sequencing could be obtained, this study employs
the respective evaluation to solve the group decision problem.
The respective evaluation: When using this individual method, the attribute
weights of criteria and alternatives applied by each member in the group can
be different. Suppose that the union of criterion sets used by the decision
maker i = 1,…, n is C = {c_{1}, c_{2}, …, c_{t}},
its weight vector is W^{i} = (w^{i}_{1}, w^{i}_{2},...,
w^{i}_{p}):
and w^{i}_{q}≥0, if the member i adopts the criterion p,
w^{i}_{p} = 0.
Table 1: 
Evaluation indexes and weights of extra curricular sports
lifestyle on college students 

Following the steps of ELECTREII, we can get the individual ranking of alternatives
by each decision maker in the group. After all the members sort the alternative
sets, the ELECTREII method can be used to gather these individual rankings
to form the group sequencing.
Firstly, suppose that the weight of the decision made by the decision maker
i = 1,…, n in the group decision is w_{i}ε{w_{1},
w_{2}, …, w_{n}} and the decision committee can obtain
the weights of decision makers by the AHP method (Mayster
et al., 1994). If the individual ranking of the member i shows that
the alternative x_{k} is better than the alternative x_{l} (x_{k}φ_{i}
x_{l}) and the set of all decision makers i satisfying the condition
x_{k}φ_{i} x_{l} is written as I^{+}(x_{k},
x_{l}), the set of members i satisfying the condition x_{k}~_{i}x_{l}
is I^{=}(x_{k}, x_{l}) and the set of members i satisfying
the condition x_{k}~x_{l} is I¯(x_{k}, x_{l}).
Calculate the harmony indexes:
Determine the high, medium and low thresholds α*, α^{0} and
α¯, 0.5<α¯<α^{0}<α*<1. Given
d^{0}_{i}<d*_{i} and define D^{h}_{i},
D^{m}_{i} and D^{l}_{i}. Define the strong outranking
relation and the weak outranking relation:
The forward strong and weak relation graphs G_{s} and G_{w}
of alternative sets were constructed by the outranking relations of all the
alternatives obtained by the above equations.
Firstly, make sort ascending by the directive diagram, calculate the sort v’(x_{i})
of each alternative and draw the ranking table. Then mirror the forward strong
and weak relation graphs to get the sort descending diagram, calculate the order
v^{0}(x_{i}) of alternatives’
sort descending by using the same method and draw the sort descending table.
Combined with the results of sort ascending and sort descending, by the equations:
v(x_{j}) = 1+v*v^{0}(x_{j}) 
(6) 
Calculate the mean sort
of alternatives and draw the mean ranking table of alternative sets and the
group can get the final sort of alternative sets by the rule that the smaller
the
is, the higher the rank of the alternative.
RESULTS AND DISCUSSION
Taking the pratical decision problem as an example, we apply the respective
evaluation to illustrate the above process and its reasonability. For ease of
analytical calculation, suppose that an evaluation committee composed of p_{1},
p_{2}, p_{3}, p_{4} and p_{5} evaluates x_{1},
x_{2}, x_{3}, x_{4} and x_{5} in the alternative
set X of three persons.
Individual decision: According to the criteria (the weight of attributes),
the committee members sort all the persons by the ELECTREII:
Step 1: 
Decision makerp_{1} gets the weighting of index attributes
by the AHP method to reflect the relative importance of all attributes.
The weight vector of the index system is set as w_{1} = {w_{11},
w_{12}, w_{13}, w_{14}, w_{15}} = {0.1,
0.2, 0.25, 0.2, 0.25}. The pairwise comparison between persons according
to the attribute y_{j}(j = 1,…, 5) can obtain the set of all
attributes y_{j} satisfying the condition x_{k}™_{j}x_{l},
the set J = (x_{k}, x_{l}) of all attributes y_{j}
satisfying the condition x_{k}~_{j}x_{l} and the
set J¯ = (x_{k}, x_{l}) of all attributes y_{j}
satisfying the condition x_{k}™_{j}x_{l}: 
Calculate the harmony indexes:
Step 2: 
A possible acceptable minimum value of resection is provided
to each attribute by the connection method, namely only when each attribute
value of an alternative is no lower than its corresponding value of resection,
the alternative may be accepted, based on which the disharmony set D_{j}
is constructed. Set the thresholds d*_{j} = 0.5, d^{0}_{j}
= 0.3, if y_{jk}y_{jl}>d_{j}, then the compensation
of the other attributes is no longer accepted, namely x_{k}™x_{l}
can not be admitted. Three disharmony sets D^{h}_{j}, D^{m}_{j}
and D^{l}_{j} can be defined according to the ‘Decision
Theories and Methods’ 
Step 3: 
Given the high, medium and low thresholds of the harmony indexes, 0.5<α¯<α^{0}<α*<1,
in which α¯ = 0.6, α^{0} = 0.8 and α* = 0.9.
Define the strong and weak outranking relations (O_{s} and O_{w}) 
Step 4: 
Based on the above strong outranking relation O_{s} and the weak
outranking relation O_{w}, respectively draw the strong relation
graph G_{s} and the weak relation graph G_{w} for the alternative
set X 
Step 5: 
Make the sort ascending by the forward directive diagram and calculate
the sort v’(x_{i}) of all alternatives, as shown in Table
1 

The sort v’(x_{i}) of all alternatives: 
Step 6: 
Mirror the strong and weak relation graphs of sort ascending
to get the sort descending diagram. Make the sort by the sort descending
diagram, calculate the sort descending value v^{0}(x_{i})
of each alternative and list out the calculated values v(x_{i})
of sort ascending and descending for each alternative, as shown in Table
2 
Calculate the mean sort (x_{i})
for each alternative as shown in Table 3. The sort of alternatives
made by decision maker p_{1} is:
The attribute weight given by the decision makers is:
Table 2: 
Experimental results of v^{0}(x_{i}) and
v(x_{i}) 

Following the same steps can obtain the individual ranking of alternatives
by the other members:
Group decision:
Step 1: 
Group determines the weight of each member by the AHP method
as W = {0.3, 0.2, 0.25, 0.15, 0.1}. The individual ranking made by the decision
maker i shows that the alternative x_{k} is better than the alternative
x_{l}(x_{k}™_{i}x_{l}). The set of
all decision makers i satisfying the condition x_{k}™_{i}x_{l}
is written as I^{+}(x_{k}, x_{l}), similarly, the
set of members i satisfying the condition x_{k}~_{i}x_{l}
is i = (x_{k}, x_{l}) and the set of members i satisfying
the condition x_{k}™_{i}x_{l} is I¯(x_{k},
x_{l}). Calculate the harmony indexes and the experimental results
can be listed as follows: 
Step 2: 
Given the high, medium and low thresholds α*, α^{0}
andα¯, 0.5<α¯<α^{0}<α*<1;
given d^{0}_{i}<d*_{i} and define D^{h}_{i},
D^{m}_{i} and D^{l}_{i}. Define the strong
and weak outranking relations based on which the strong relation graph G_{S}
and the weak relation graph G_{W} are constructed 
Step 3: 
Make the sort ascending by the forward directive diagram and calculate
the sort v’(x_{i}) of all alternatives, as shown in Table
4 
Step 4: 
Mirror the strong and weak relation graphs of sort ascending to get the
sort descending diagram. Make the sort by the sort descending diagram, calculate
the sort descending value v^{0}(x_{i}) of each alternative
and list out the calculated values v(x_{i}) of sort ascending and
descending for each alternative as shown in Table 5 
Table 3: 
Experimental results of (x_{i}) 

Table 4: 
Experimental results of v’(x_{i}) 

Table 5: 
Experimental results of v^{0}(x_{i}) and
v(x_{i}) 

Table 6: 
Experimental results of (x_{i}) 

Calculate the mean sort (x_{i})
for each alternative as shown in Table 6. Thus, the sort of
alternatives made by the group is:
CONCLUSION
The indexes were determined by the Delphi method. After expert surveys, the
AHP was employed to analyze the results, obtaining the weights and subweights
of all indexes. The grade evaluation criteria for indexes of extracurricular
sports lifestyle on college students were built by the percentile method. The
extracurricular sports lifestyle of 100 students in Wuhan University of Technology
was tested by selfassessment and expert judgment make statistics of the results
and compare them with the results from expert judgment. Upon testing, there
is a significant relation between the evaluation result and the result of expert
evaluation (p<0.01), indicating the consistency of the results obtained from
these two evaluation methods is good.
The analysis of reliability and validity of evaluation results: The selfassessment
reliability coefficient is 0.754, the expert evaluation reliability coefficient
is 0.769, so the reliability is high and the evaluation results are reliable;
the overall validity coefficient is 0.968, so the coefficient error is small
and the validity is high. Thus, the evaluation index system of extracurricular
sports lifestyle on college students is of high reliability and validity.