An

Examination of the Effects of Economic

and In-Game Variables on National
Basketball Association Stadium
Performance and Fanbase Loyalty
Nicholas Heath

ABSTRACT
The sports economics literature identifies a set of variables that affect
ticket pricing at sporting venues. This paper revisits these variables in the
National Basketball Association and finds a new significant contributor to
pricing, the geographic distance between city centers and stadiums.
Additionally, using a random-effects model with panel data, I evaluate
variables affecting stadium attendance and show significant correlative
ticket price and team performance interplay. Finally, I create a fan loyalty
index via win percentage and attendance correlations as a proxy and
examine these relationships across various city economic factors.
1.1 Introduction

The modern day sports industry acts as the crossroads between two major entities:

the game itself, and the economic and business factors that make the industry viable in

the first place. Over the last half century, the game of basketball has evolved

tremendously at the professional level, and along with it has grown the

commercialization of the industry. After Michael Jordan popularized and sensationalized

the league in the 1990s, the NBA rapidly grew into the 5.2 billion dollar per year industry

it is today.1 Constituting a majority of that revenue, even above the leagues new record

TV deal, is the cash flow from filling in the seats at venues across the country night after

night.2

This paper will examine the factors affecting attendance at NBA venues and the

interplay between in-game and economic contributors to attendance. I will first develop a

model for determining ticket pricing using linear regression. Various prior studies have

created models across the major US sports leagues for estimating attendance. We will use

similar methodology to develop a model for setting ticket prices for a given team. This

model will depend on a host of variables, both pertinent to economic conditions of the

city and to recent and current performance of the team. Controlling for city size, per

capita income, and prior performance of the team, we will assess the largest contributors

to pricing variance. Due to limited availability of data, this first analysis will focus

exclusively on current season (2015-2016) parameters. Given the dearth of sample size (n

= 30 because the league itself only has 30 teams), significance was difficult to prove.

However, controlling for median income per city, there was a significant relationship

1
Statista.com
2
Moore
3
Rishe and Mondello, 2003
4
Rishe and Mondello, 2004
2
between population and ticket price and geographical distance between the stadium and

the city center. Although there is no demonstration of causality, prior literature and

economic theory supports this finding.

After this evaluation of ticket price variance, I will do an in depth team-by-team

exploration of the interplay between team performance and attendance. That is, I will

survey the effect win records have on fan base tendency to fill out the stadium. This

relationship could potentially serve as a proxy for the loyalty of the teams fanbase: as

teams alter their performance, do their fanbases alter their attendance behavior or do they

remain loyal? As predicted, the strength of this tie will be highly contingent upon the

economic factors of the city in which the team plays. That is, given the prior analysis of

ticket pricing, it should follow that the attendance of teams playing in wealthy cities with

large populations will depend far less on the actual performance of the team. In these

instances of large market teams, such economic factors should effectively drown out

much of the variation one would expect to see from variation in performance.

1.2 Background

A collection of studies examining the relationship between ticket pricing and team

effects across various American sports has highlighted potentially important factors. For

simple supply-demand economics, the expected effects of population and income

variables are relatively straightforward. Increases in both should expand the demand

curve and raise prices. Furthermore, superior team performance should encourage

attendance across all portions of the demand curve. Consistent in all of these studies is a

highly significant effect of population on pricing, as is expected. One study, examining

NFL pricing data from 1996 to 2001, also finds a significant relationship between prior

3
performance and price and between family income and price (Rishe and Mondello,

2003). 3 Another study, spanning all four major US sports leagues (NBA, NHL, MLB,

NFL) from 1996 to 2002 finds that, in addition to population effects, moving to a new

stadium causes a significant increase in pricing.4 I will examine all these variables and

others to develop a ticket pricing model. I will also introduce a set of variables

unexplored in previous literature, namely team popularity and name recognition and

distance between the stadium and the center of the city (termed geography effects). One

potential variable, stadium concession pricing, proved to be a significant contributor of

ticket price in the MLB, but unfortunately no such comprehensive data on NBA

concessions pricing exists.5

The main divergence from prior literature, however, will be the creation of the

team loyalty index. That is, as explained prior, the correlation between team

performance and attendance. These loyalty calculations can be thought of as elasticity

of attendance, as they measure the sensitivity of attendance to win percentage. The

correlations alone will not control for other economic factors, and so will allow us to

observe how they relate to city population and median income.

2.1 Data, and Endogeneity Concerns

Collection of the data was a highly manual process, involving the screening of a

variety of different informatics sites. Given the necessity to assemble the data

independently, a host of limitations arise.

3
Rishe and Mondello, 2003
4
Rishe and Mondello, 2004
5
Krautmann and Berri

4
 Ticket pricing was provided by a comprehensive data set put together by Team

Marketing. The price per team by year was concluded by averaging the presale value of

every ticket in the stadium. That is, for a given stadium, it takes the average of seats

ranging from nosebleeds (far from the court, low value seating) to courtside (high value)

seats. It excludes box seating and factors in season ticket pricing. Thus, average ticket

prices for a given team are susceptible to how the stadium sells and determines the

breakdown of its high and low value seating. Furthermore, the prices listed are explicitly

set by teams pre-season, and thus do not necessarily reflect free market resale and

valuation as determined by ticket resale services (such as StubHub and TicketMaster, the

secondary ticket market). The pricing goes back to 1992 and includes up to the 2013-

2014 season. Notably, the ticket prices are not inflation-adjusted. To make this

adjustment, instead of using inflation rates over the last 25 years, I looked at the

inflation rates of the tickets themselves. That is, I looked at the increase in average

ticket price of all teams throughout the league by year and discounted all prices by this

increase by year.

Individual team performance by year is available from the Basketball-

Reference.com team index. 6 Although no data set existed showing year-by-year

performances for all 30 NBA teams, I was able to find historical performance for

individual teams and compile them together into a new data set. Although team

performance is available for a select few teams from the leagues founding in 1946 up

until the most recent, I collected data only from 1992-2014 in accordance with available

6
http://www.basketball-reference.com/teams/

5
ticket pricing data. Team performance is calculated as regular season win-loss record, and

does not include playoff performance.

Stadium attendance data was collected from the Association for Professional

Basketball Research.7 The data span all years of teams operations, and are presented as

raw attendance totals averaged over all home games for a given season. For example, in

the 2005-2006 season, the Boston Celtics seated a total of 692, 873 attendants. Over their

41 home games that season, they averaged 16,899 attendants per game. To normalize for

varying stadium size, I divided the attendances by the total seats in each stadium, which I

found from the Official National Basketball Association Guide 2014-2015.8

An important factor that arose during my data collection was the relocation of

teams and construction of new teams. A majority of NBA teams have either switched

cities or moved to new stadiums within their home city since the beginning of the data

collection starting in 1992. In fact, only 6 teams have consistently played in the same

stadium since 1992 (the Detroit Pistons, Milwaukee Bucks, Minnesota Timberwolves,

New York Knicks, Sacramento Kings, and Utah Jazz). Given the massive implications on

attendance and ticket pricing that arise from a teams changing stadium or cities, I

segmented attendance, pricing, and performance data by current teams most recent

relocations. For example, the Chicago Bulls moved in 1994 from the Chicago Stadium to

the newer, larger United Center. Following this move, the attendance numbers, as

expected, spiked up by over 4,000 fans per game to fill in the larger stadiums seating.

Thus, I went through the team history of all current NBA teams to find out when their

most recent relocation occurred, and retained only data from these relocations.

7
http://www.apbr.org/attendance.html
8
https://mediacentral.nba.com/media/mediacentral/2014-NBA-Guide.pdf

6
Consequently, due to a small sample size of years, I dropped two teams from my

analysis, the Orlando Magic, who moved to a new stadium in 2010, and the Brooklyn

Nets, who moved from New Jersey to Brooklyn in 2012. The team with the most recent

relocation retained is the Charlotte Hornets, who moved to a new stadium in 2005. Thus

only 9 years, a relatively small sample size, were collected for this team, as were for a

host of other teams. These limitations will decrease certainty around the results found. As

an additional complicating factor, there exists a Honeymoon Effect with new stadiums,

wherein the construction and relocation to a new stadium causes a temporary spike in

attendance and pricing, as noted by Zygmont and Leadley, 2005.9 Thus, my modeling has

no control for this attendance effect.

For the ticket pricing model, I collected a host of other data potentially relevant to

ticket pricing. First, I found the population data from 2014 as estimated by the US

Census. 10 Limitations within year-by-year estimations prevented the inclusion of

population in the later time series. Per capita income was also retrieved from the US

Census as measured in 2010. Thus, it must be noted that the income data might be

slightly dated. For the Toronto Raptors I had to consort a Canadian city database and

adjust for the exchange rate for the Canadian and US dollar. As a proxy for the teams

popularity and name recognition, I marked down the number of Likes on each teams

Facebook page. The number of Likes likely to be contingent upon the market size of

the team itself, thus potentially limiting avenues for causality. However, take, for

example, the Los Angeles Lakers and Los Angeles Clippers, who play in the same

stadium in the same city with the same market size. The Lakers have 21 million

9
Leadley and Zygmont
10
http://factfinder.census.gov/faces/tableservices/jsf/pages/productview.xhtml?src=bkmk

7
Facebook Likes vs. the Clippers 3 million, likely due to the prolific history of the Lakers

coupled with the superstars who have cycled through Lakers rosters (Kobe Bryant,

Shaquille ONeal, Magic Johnson, Kareem Abdul-Jabbar), as is described by the

superstar effect (Berri and Schmidt, 2006).11

Finally, I looked at a variable never before examined in sports ticket pricing

literature, the distance between the stadium and the downtown city center. To collect this

data, I used Google Maps directions for all thirty teams to calculate the driving distance

necessary to attend games. As it follows, large distances between the center of the city

should adversely affect fans propensity to attend these games and, subsequently, should

lower ticket pricing due to the decrease in demand. The most prominent example comes

from the Detroit Pistons, who play at The Palace at Auburn Hills, located a full 32 miles

from downtown Detroit. As will be seen later, this distance likely plays a significant role

in fan attendance elasticity.

Thus, with the preceding data, I assembled two data sets. First, a freeze-frame of a

host of variables concurrently affecting ticket prices. Second, a panel data set

demonstrating team by team attendance, ticket pricing, and win percentage over 22 years

(or for as long as the team has existed in its current location).

2.2 Methodology

Given that prices are the averages of pre-season ticket sales, the first data set uses

the winning percentage and attendance from the prior season, as these are the only

variables available for the teams ticket-selling purposes. Similarly, the expected effect

of win percentage on attendance is not completely time-independent. That is, the win

11
Berri and Schmidt

8
percentage is determined as the season progresses, as is the attendance. Thus, attendance

responds to win percentage game by game. However, linked with win percentage likely

exists the omitted variable of pre-season and mid season forecasting. That is, fans dont

necessarily assess how good a team is exclusively through their win loss record. They

might be encouraged, instead, by the high level of play expected from their team or

potentially by the excitement and media focused generated from a their teams success.

Thus, the model is as such:

TICKET_PRICE = 1ATTENDANCE2016 + 2POPULATION +

3FACEBOOK_LIKES + 4MEDIAN_INCOME + 5WIN%2015 +

6MILES_DOWNTOWN + 7POP_DIST 8POP_INC + 10 + ERROR

Two interaction terms are introduced in the model. First, population and distance

could potentially be interactive given that there will be more fans in cities with densely

populated city centers who will have to travel. Second, income might affect the

population effects seen since an increased population will affect seat filling contingent

upon these extra fans being able to actually afford seating, especially expensive seating,

to bid up the market price.

3.1 Linear Regression Results

After running the regression, only two variables showed significance. First, in

accordance with virtually all prior pricing literature, population seems to be strongly

correlated with ticket price. For every increase of one million inhabitants in the city,

ticket price increases by about 5 dollars, or a 1% increase. This is noticeable less than the

3% increase observed by Riche and Mondello. This difference could be due to their use

9
of non-census population data (their population was retrieved from dismal.com), their use

of panel data, or the fact that their time covered is the late 1990s and early 2000s as

Table 1: Linear Regression with Ticket Pricing as the Dependent Variable

Average Ticket Price Coefficient Standard Error

Attendance 2016 -0.555 0.665
Population 5.07* 2.02
Facebook Likes (Millions) 1.435 0.781
Median Household Income -0.000145 0.000642
Win % 2015 -55.826 138.163
Miles from Downtown -0.315* 0.166
Population_Distance -2.53 1.68
Population_Income 0.00315 0.00219
Constant 77.315 64.714

Figure 1: Average Ticket Price vs.

Population
140
120
Average 100
Ticket 80
Price ($) 60
40
20
0
0 1 2 3 4 5 6 7 8 9
Population (Millions)

opposed to the mid 2010s for these data. Additionally, distance between the stadium and

the city center showed a statistically significant negative relationship. For every extra

mile in driving distance between the downtown center of the city, ticket price drops by 31

cents. No other regressors, including the interaction terms, showed significance.

10
3.2 Random Effects Model Results

The panel data regressions run use the random effects, or various components

model, wherein entity fixed effects per team are not included due to demographics and

institutional changes in team behavior. For example, the quality of coach affects win

percentage, and coaches are often traded, fired, and hired among teams. Similarly,

players are traded from team to team frequently, which can lead to sporadic fluctuations

in culture and chemistry. Therefore, capturing these dynamic effects was prioritized over

controlling for time invariant fixed effects, which are expected to be small due to erratic

team behavior. Using fixed effects would have assumed autocorrelation within entity

observations, and because of the small correlation between the same team over time due

to changing exogenous factors, this assumption would have been invalid.

The results from the panel data showed high significance. On average, a one

percentage point increase in win percentage is associated with an increase of 5,891

attendants. This positive association, as outlined in the introduction, is expected to

increase the demand for tickets, subsequently raising prices. Additionally, an increase of

one dollar in price indicated an increase of about 23 attendants. This counterintuitive

result implies the possibility of omitted variable bias, wherein higher demand causes

higher attendance, and by extension, ticket prices as well. The factor driving prices and

attendance is demand, and some of demands effect on attendance spills onto the

coefficient price, skewing it upwards of its true coefficient. While seemingly

counterintuitive, this result may actually be consistent with ticket pricing practices. The

nave interpretation of this result would suggest that more people show up as a result of

higher ticket prices. However, given the causality concerns that arise from the ticket data

11
coming from pre-sale as set by the stadium, it actually follows that increased ticket price

would have such a small coefficient (23 additional attendants is about a .1% increase in

total attendants). This is because those who set the ticket prices may do so with the intent

of maintaining total attendance numbers at a predetermined, profit-maximizing quantity.

Thus, ticket sales would be adjusted by the price setters to keep attendance constant. The

low significance and high p value of this regressor (.03) may act as evidence for this

phenomenon. Issues about understated standard errors may arise; however, to inhibit such

bias, standard errors are clustered by team and are therefore larger, which prevents type 2

errors in wrongly accepting statistical significance.

Table 2: Random Effects Model with Attendance as the Dependent Variable

Attendance Coefficient Standard Error

Win Percentage 5891.063** 820.813
Ticket Price 23.126* 10.746

3.3 Attendance Elasticity Indexing

Finally, the team by team calculations of attendance elasticity were very

interesting. These correlations, sorted from weakest to strongest, are shown below:

Table 3: Attendance Elasticity by Team

Team Loyalty, Attendance Elasticity (Win%-Attendance Correlation)

Dallas Mavericks -0.0363
New Orleans Pelicans 0.1182
New York Knicks 0.2142
Charlotte Hornets 0.2799
Minnesota Timberwolves 0.2816
Atlanta Hawks 0.2944
San Antonio Spurs 0.3873
Milwaukee Bucks 0.4065
Boston Celtics 0.4385
Los Angeles Lakers 0.4392

12
 Miami Heat 0.4853
Indiana Pacers 0.5549
Toronto Raptors 0.5571
Washington Wizards 0.5751
Phoenix Suns 0.6037
Sacramento Kings 0.6216
Chicago Bulls 0.6332
Cleveland Cavaliers 0.6425
Detroit Pistons 0.6732
Philadelphia 76ers 0.6822
Golden State Warriors 0.6953
Portland Trailblazers 0.7076
Memphis Grizzlies 0.7493
Utah Jazz 0.7718
Houston Rockets 0.787
Orlando Magic 0.8326
Denver Nuggets 0.8367
Los Angeles Clippers 0.8393

Correlations are drawn from as far back as the teams most recent relocation. Thus, some

sample sizes may be relatively small. Furthermore, given teams differing relocation

times, sample sizes vary across teams. To visually represent how this metric demonstrates

fanbase loyalty, I selected two disloyal fanbases (the Portland Trailblazers and Detroit

Pistons, showing correlation values of .7076 and .6732, respectively) and two loyal

fanbases (the Dallas Mavericks and New York Knicks, with correlations of -.0363 and

.2142, respectively).12 First, the disloyal fanbases:

12
The negative correlation of Dallass win percentage and attendance shows that these
two variables were so independent that increases in win percentage were actually
accompanied, counter intuitively, by decreases in attendance.

13
 Figure 2: Portland Trailblazers Win
% vs. Attendance %
1.2

0.8

0.6 Attendance Percentage

0.4 Win Percentage

0.2

0
1996 2000 2004 2008 2012

Figure 3: Detroit Pistons Win % vs.

Attendance %
1.2

0.8

0.6 Attendance Percentage

0.4 Win Percentage

0.2

The visualizations are striking; while the graphs alone do not demonstrate

causality, it is clear for these two teams that attendance percentage very closely follows

the teams win percentage. As the Trailblazers performance dipped in the mid 2000s, so

did their fanbase disappear from the stadium. Shortly thereafter, however, their

14
performance picked up and the fans returned to the stadium. One potential interpretation,

the one that links these correlations to disloyalty, is that Portland fans only support their

team when it is doing well and do not stick by their team when it is underperforming. The

Detroit Pistons representation shows a more one-sided story: as the Pistons mid 2000s

success rapidly steeped off, their fans quickly responded by staying home and leaving

over one third of the arena empty. As noted earlier, this effect may be confounded by the

Pistons stadiums exceptionally far distance from central Detroit.

Conversely, the Knicks and the Mavericks show an alternative case, that of the

diehard fanbase, which doesnt waver as a result of team performance fluctuations:

Figure 4: Dallas Mavericks Win % vs.

Attendance %
1.2
1
0.8
0.6 Attendance Percentage
0.4 Win Percentage
0.2
0

15
 Figure 5: New York Knicks Win % vs.
Attendance %
1.2
1
0.8
0.6 Attendance Percentage
0.4 Win Percentage
0.2
0

Attendance percentage remains completely independent of win percentage, regardless of

how well the team actually does. In contrast with the high correlation teams, once again

using the loyalty explanation, one would postulate that fans are less partial to their teams

performance in deciding whether to attend games and will show up to the stadium

regardless.

Of course, this loyalty metric quite possibly does not actually measure the

loyalty that fans hold for their teams. While innate propensity to attend games may be

part of these decisions, a set of other factors is likely contributing to fan willingness to go

to games. To tease out and help hypothesize as to the potential causes, I look at the teams

that have representations similar to those of the Knicks and Mavericks, wherein

attendance remains at or near full capacity, even when, as is the case with the Knicks, the

team is at the bottom of the league. The simple intuition there is that in a city like New

York, given its high population and median income, demand for attendance will be so

tremendously high that it will effectively drown out whatever attendance effect win

percentage may have.

16
 Thus, with little probing for causality, I scrutinize this seat-filling effect by

examining how correlation values differ in high income and high population cities vs.

low income and low population cities. High and low are assessed by taking the top

half or bottom half of the most populous or highest income teams.

As predicted, increased population and increased median income are met with

much lower correlation values. Teams in high income cities have an average correlation

of .4200 whereas those in low income cities have a correlation of .5893. Similarly, teams

in high population cities have a correlation of .4140 vs. low population at .5514. It must

be noted that these correlations are not averages of the loyalty index shown earlier but are

R values collected over the time series containing 449 observations.

4.1 Limitations and Discussion

In addition to the limitations addressed in the data section earlier, there arise a set

of other problems when interpreting the data. These stem from lack of causality and

omitted variable bias. For example, if the decision to prioritize uncorrelated standard

errors due to erratic team behavior to use a random effects model was incorrect, the

absence of fixed effects introduces omitted variable bias that incorrectly skews the

coefficients. In that case, one should maintain cautionary interpretation of the coefficients

in the regression. However, constantly changing team behavior in terms of relocation,

sponsorship, player movements, team advertising etc. still indicates validity in choosing

GLS random effects.

First, virtually none of the relationships shown actually demonstrate causal

mechanisms. For example, although population correlates strongly with increased ticket

pricing, I was not able to find any instrumental variables that could show that population

17
explicitly was causing this increase. For example, increased population could correlate

with increased number of corporate ticketholders in the region that bid up the pricing,

which could potentially explain why the New York Knicks attendance is so

unresponsive to win percentage. Similarly, in the panel data regression, there were no

valid instruments found. Thus, win percentage alone may not be causing attendance

fluctuations; for example, media hype surrounding winning teams might act as a

marketing tool that exogenously shifts out the demand curve for games. Furthermore,

there exists the possibility for reverse causality. Attendance may be driving win

percentage; although actual attendance fluctuations relative to the size of the stadium tend

to be small, it may be the case that teams perform better when attendance is higher. The

home advantage is well documented within the NBA, with the home team winning about

61% of games, the highest of all US major sports leagues. Furthermore, part of the

causation of this advantage likely lies within the support that the home crowd plays.13

Thus, it is possible that part of the reason we see this positive correlation is because

teams play better when supported by a larger crowd.

While the final observations regarding loyalty differences between high and low

income cities are interesting, it is imperative that we interpret them for what they are. The

mere fact that high income, high population teams tend to exhibit less of a relationship

between attendance and win percentage does not necessarily mean that it is because of

these economic factors that fans are or not filling up the seats. They could be doing so

due to a host of omitted variables, or simply because, as explained earlier, actual fan

loyalty itself might be contingent upon the market in which the team plays.

13
http://stat.wharton.upenn.edu/~dsmall/nba_rest_submitted.pdf

18
 Despite internal validity concerns, various extrapolations can be drawn from the

above results. Foremost we draw greater insight into the important factors that help

determine successful stadium and city relocations and expansion teams, noting that, as

expected, higher population and minimized distance from city center are significant

drivers of ticket pricing. Additionally, proper team-by-team assessment of their loyalty

indices may be useful for determining the team-specific factors that can be used to best

draw out fanbase attendance.

4.2 Further Research

The results shown open up a host of possibilities for future research. Foremost is

the search for instrumental variables to demonstrate causality. Although I found none,

there likely exist regressors that might show potential causal relationships. For example,

although I was not able to find a comprehensive, team by team data set showing

aggregate player payroll, it is quite possible that this parameter might correlate with win

percentage but not with attendance, as payroll reflects the talent on the team but wouldnt

necessarily draw out larger crowds. Assuming these theoretical premises hold, one may

argue player payroll is a plausible instrument.

Additionally, these methods could be applied to other major US sports leagues. It

could be interesting to develop attendance elasticities not for teams but for entire leagues.

This could reveal certain aspects regarding the culture surrounding the fanbases of these

sports. For example, its possible that baseball may have a lower attendance elasticity

than basketball since attendance of baseball games may be more predicated on enjoying

the experience as a whole and not necessarily focusing on the game itself. Furthermore, if

proper controls for city economic variables were taken into account, it could be possible

19
to demonstrate if fanbases of certain sports are more likely to be involved in their sport.

For example, hockey is often considered to have the most diehard following; would this

be reflected in lower attendance elastic than the other major leagues?

5. Conclusion

This paper thus achieves a host of goals. First, in examining ticket price

contributors, it confirms prior literature in other leagues finding that population has a

significant effect on ticket price, and contributes another significant variable, distance

from downtown, to the discussion. Second, it shows strong correlative relationships

between attendance, win percentage, and ticket pricing across all 30 teams over a 22 year

span. Finally, it develops a new, potentially interesting parameter, elasticity of

attendance, which could potentially act as a proxy for fanbase loyalty, and shows that

decreased loyalty is exhibited in lower-population, lower-income cities.

20
 Works Cited

"2015-16 Official NBA Guide and NBA Register." NBA.com. N.p., n.d. Web. 08 May

2016.

Berri, David, and Martin Schmidt. "On the Road With the National Basketball

Association's Superstar Externality." Journal of Sports Economics (n.d.): n. pag.

2006. Web. 8 May 2016.

Krautmann, Anthony, and David Berri. "Can We Find It at the Concessions?

Understanding Price Elasticity in Professional Sports." Can We Find It at the

Concessions? Understanding Price Elasticity in Professional Sports. Journal of

Sports Economics, n.d. Web. 08 May 2016.

Leadley, John, and Zenon Zygmont. "When Is the Honeymoon Over? National

Basketball Association Attendance 1971-2000." When Is the Honeymoon Over?

National Basketball Association Attendance 1971-2000. Journal of Sports

Economics, May 2005. Web. 08 May 2016.

Moore, Matt. "Report: NBA Set to Announce New Television Deal worth $24 Billion."

CBSSports.com. CBS Sports, 5 Oct. 2014. Web. 08 May 2016.

"NBA & ABA Team Index | Basketball-Reference.com." Basketball-Reference.com.

N.p., n.d. Web. 08 May 2016.

"NBA Home Attendance Totals." NBA Home Attendance Totals. N.p., n.d. Web. 08 May

2016.

Rishe, Patrick, and Michael Mondello. "Ticket Price Determination in Professional

Sports: An Empirical Analysis of the NBA, NFL, NHL, and Major League

Baseball." Ticket Price Determination in Professional Sports: An Empirical

21
 Analysis of the NBA, NFL, NHL, and Major League Baseball. N.p., n.d. Web. 08

May 2016.

Rishe, Patrick, and Michael Mondello. "Ticket Price Determination in the National

Football League: A Quantitative Approach." Ebcohost.com. Sport MArketing

Quarterly, 2003. Web. 8 May 2016.

"Topic: National Basketball Association (NBA)." Www.statista.com. Statista, n.d. Web.

08 May 2016.

"Your Geography Selections." American FactFinder. N.p., n.d. Web. 08 May 2016.

22