FLAGS (EGM)Preliminary Results of Phase IV: development of the FocaL Adult

Gambling Screen

Tony Schellinck, Focal Research and Dalhousie University

Tracy Schrans, President of Focal Research

Michael Bliemel , Dalhousie University

Heather Schellinck, Dalhousie University

8th European Conference on Gambling Studies and Policy Issues Sept 2010

Research funded by Ontario Problem Gambling Research Centre – still under review1© Focal Research

Our Vision

FLAGS (EGM) is the next generation of measurement for adult gambling risk and harm specifically designed for public health surveillance. FLAGS moves beyond traditional identification of problem gambling prevalence providing information for use in informing, monitoring and evaluating gambling related prevention, harm reduction, social policy and public health policy for youth.

2© Focal Research

Our Purpose: Public Health Surveillance

• According to the World Health Organization (WHO) Public Health Surveillance is “the ongoing, systematic collection, analysis and interpretation of health-related data essential to the planning, implementation, and evaluation of public health practice”

• The primary goals of surveillance as defined by WHO are:

– To serve as an early warning system - measure risk

– Guide public health policy and strategies – Provide actionable information

– Document the impact of an intervention or progress towards specified public health targets/goals – Be highly accurate and consistent

– Understand and monitor the epidemiology of a condition in order to set priorities and guide public health policy and strategies – be causally related and detailed enough for policy development

• http://www.who.int/immunization_monitoring/burden/routine_surveillance/en/index.html3© Focal Research

Why is FLAGS (EGM) needed? • Current screens are probably reasonably accurate in identifying

adults who are suffering more severely from the effects of gambling. However, these instruments were not designed to identify gamblers who may be at different risk levels or to provide detailed information to the decision maker. Moreover, the goal is to use surveillance to inform, monitor and evaluate it is equally important to know ‘why’ a person is at risk especially if the goal is to help them.

• FLAGS is designed to do just that. FLAGS also more accurately assigns adults to a problem gambling category as it has a larger inventory of consequences than other screens as well as a measure of persistence that helps to overcome the transient nature of adult behaviour and properly define a problem gambler.

• Because all of the constructs are multi-item measures that are reliable and valid, the FLAGS instrument allows for accurate comparisons among jurisdictions of the prevalence of problem gambling and perhaps more importantly the extent to which the existing adult populations are at risk of developing problem gambling.

4© Focal Research

What is the FocaL Adult Gambling Screen (FLAGS)?

• FLAGS (EGM) is a new survey instrument designed to gather adult EGM gambling information in order to inform social and public health policy.

• ≈ 58 statements • Ten Constructs that provide indications of risk or problem

gambling• A scoring system that allots people to one of five categories

based their being flagged on a combination of indicators• A reporting system - set of predesigned analyses that will

provide information necessary for decision makers– Validity and reliability checks of the constructs and their

relationships in the setting it was administered– The prevalence and profiles of the each risk category

5© Focal Research

FLAGS Five Levels of Player Risk for Machine Gambling

Risk Level Label Description

Level V Problem Gambler A Problem Gambler is a person who flagged as exhibiting both

Negative Consequences and Persistence.

Level IV Advanced Risk Those at Advanced Risk are not flagged as a problem gambler

but have one or more indications on the five constructs

ranked highest in the hierarchy score: Impaired Control:

Begin, Preoccupation: Obsessed, Risky Practices: Later,

Negative Consequences and Persistence.

Level III Intermediate Risk Those at Intermediate Risk are not Problem or Advanced Risk

gamblers, but have been flagged on one or more of the

intermediate risk constructs: Impaired Control: Continue and

Risky Practices: Earlier.

Level II Early RiskThose at Early Risk have flagged on at least one of Risky

Cognitions: Beliefs, Risky Cognitions: Motives or

Preoccupation: Desire.

Level I No Risk Those at No Risk do not flag on any of the risk indicators

although it is possible that they answered yes to one or more

statements making up some of the constructs.

Level 0 Non-gambler A Non-Gambler by definition is at no-risk because they are

not involved in the activity over the last year. 6© Focal Research

How was the FLAGS developed and tested?

Built on earlier work, 1998 NS VLT study, 2000 Follow-up study, SAPGS development for the Victoria Gambling Research Panel (2004) and the quantitative testing phase for the Victoria Department of Justice (2006).

Phase I – literature review of method bias in measures– funded by OPGRC

Phase II – Literature review of Formative versus Reflective constructs – funded by OPGRC

Phase III – qualitative and preliminary quantitative research on 190 statements and constructs that were eventually rejected (e.g., superstitions) with 63 casino gamblers in Ontario – 2008 – funded by OPGRC

Phase IV 2009 - 10 quantitative analysis of responses to 132 statements to formulate FLAGS (EGM) beta – 384 Ontario Casino Gamblers – Funded by OPGRC

Phase V 2010 - Further quantitative refinement of instrument and testing on VLT gamblers – 500 VLT gamblers – funded by Nova Scotia Gambling Foundation – report due End of September 2010

7© Focal Research

Risky Cognitions:

BeliefsFormative

Risky Cognitions:

MotivesFormative

Preoccupation: Desire

Reflective

Impaired Control: ContinueReflective

Impaired Control: Begin

Reflective

Risky Practices: Later

Formative

Preoccupation: ObsessionReflective

Negative Consequences

Formative

PersistenceReflective

Risky Practices: Earlier

Formative

Formative and Reflective Constructs: FLAGS

8© Focal Research

Risky Cognitions:

Beliefs0.000

Risky Cognitions:

Motives0.243

Preoccupation: Desire0.150

Impaired Control: Continue

0.407

Impaired Control: Begin

0.423

Risky Practices: Later0.619

Preoccupation: Obsession

0.364

Negative Consequences

0.633

Persistence0.718

0.4937.70

0.4526.52

0.2914.18

0.1562.64

0.2512.92

0.2113.20

0.5699.21

0.3794.86

0.1973.76

0.4946.56

0.2383.44

0.3887.40

0.4848.14

Risky Practices: Earlier0.638

0.2053.92

0.55811.20

0.63817.20

0.2624.62

0.2863.97

PLS Model of FLAGSCoefficients, t scores and

construct variance explained

9© Focal Research

Early RiskConstructs

Intermediate Risk

Constructs

Advanced Risk Constructs Problem Gambler

Risky Cognitions:

BeliefsFormative

Risky Cognitions:

MotivesFormative

Preoccupation: Desire

Reflective

Impaired Control: ContinueReflective

Impaired Control: Begin

Reflective

Risky Practices: Later

Formative

Preoccupation: ObsessionReflective

Negative Consequences

Formative

PersistenceReflective

Risky Practices: Earlier

Formative

Risk Category Assignment: FLAGS

10© Focal Research

Median Hierarchy ranking of ConstructsMeasure developed based on Toce-Gerstein, et al. (2003)

Constructs

Median

Hierarchy

Ranking

Risky Cognitions: Beliefs 20

Preoccupation: Desire 28

Risky Practices: Earlier 45.5

Impaired Control: Continue 59

Risky Cognitions: Motives (Money

and Depression)63.5

Preoccupation: Obsessed 85

Impaired Control: Begin 104

Risky Practices: Later 103

Persistence 103.5

Negative Consequences 114

Early Risk Constructs

Intermediate Risk Constructs

Advanced Risk Constructs

Problem GamblerConstructs

11© Focal Research

Risk Profile by Risk CategoryConstructs All Gamblers No Risk

Early

Risk

Intermediate

Risk

Advanced

Risk

Problem

Gambler

Persistence 11.0% 0.0% 0.0% 0.0% 33.3% 100%

Negative Consequences 9.9% 0.0% 0.0% 0.0% 22.2% 100%

Preoccupation: Obsession 3.7% 0.0% 0.0% 0.0% 14.3% 41.4%

Impaired Control: Begin 8.6% 0.0% 0.0% 0.0% 36.1% 65.5%

Risky Practices: Later 12.8% 0.0% 0.0%% 0.0% 66.7% 82.8%Impaired Control:

Continue23.5% 0.0% 0.0% 77.3% 69.4% 100.0%

Risky Practices: Earlier 17.9% 0.0% 0.0% 47.7% 55.6% 89.7%

Risky Cognitions: Motives 17.1% 0.0% 43.6% 20.5% 38.9% 82.8%

Preoccupation: Desire 18.4% 0.0% 28.2% 45.5% 55.6% 62.1%

Risky Cognitions: Beliefs 8.8% 0.0% 35.9% 4.5% 22.2% 31.0%

12© Focal Research

Risk Profile by Risk CategoryConstructs All Gamblers No Risk

Early

Risk

Intermediate

Risk

Advanced

Risk

Problem

Gambler

Persistence 11.0% 0.0% 0.0% 0.0% 33.3% 100%

Negative Consequences 9.9% 0.0% 0.0% 0.0% 22.2% 100%

Preoccupation: Obsession 3.7% 0.0% 0.0% 0.0% 14.3% 41.4%

Impaired Control: Begin 8.6% 0.0% 0.0% 0.0% 36.1% 65.5%

Risky Practices: Later 12.8% 0.0% 0.0%% 0.0% 66.7% 82.8%Impaired Control:

Continue23.5% 0.0% 0.0% 77.3% 69.4% 100.0%

Risky Practices: Earlier 17.9% 0.0% 0.0% 47.7% 55.6% 89.7%

Risky Cognitions: Motives 17.1% 0.0% 43.6% 20.5% 38.9% 82.8%

Preoccupation: Desire 18.4% 0.0% 28.2% 45.5% 55.6% 62.1%

Risky Cognitions: Beliefs 8.8% 0.0% 35.9% 4.5% 22.2% 31.0%

13© Focal Research

Comparison to PGSIP

GSI

Cat

ego

rie

sFLAGS Categories

No RiskEarly

Risk

Inter-mediate

Risk

Advanced

Risk

Problem

GamblerTotal

No Risk 47.1% 6.4% 0.8% 0.5% 0.0% 54.8%

Low Risk 11.2% 2.4% 4.3% 1.1% 0.3% 19.3%

Medium

Risk2.1% 1.6% 6.7% 7.2% 2.7% 20.3%

Problem

Gambler0.0% 0.0% 0.0% 0.8% 4.8% 5.6%

Total 60.4% 10.4% 11.8% 9.6% 7.8% 100.0%

Overall the FLAGS categories correlation with the PGSI categories (Spearman=0.730). Risk category only correlation is 0.644 (n= 345).If it is assumed that the FLAGS Intermediate and Advanced Risk segments are roughly equivalent to the PGSI Moderate Risk there was 68.2% agreement .For those classified at some level of risk (excluding Problem Gambler) by either instrument the two instruments agreed on only 39.9% of the cases .

14

© Focal Research

What are the next steps? FLAGS

• Focal has completed a survey of 500 regular VLT gamblers in Nova Scotia for the Nova Scotia Gambling Foundation. These data have provided further testing of the FLAGS instruments on a different population and a different form of EGM gambling.

• The general report should be released by the end of this month.

• A detailed report focus on FLAGS will follow.

15© Focal Research

Cut-Offs Chosen for IndicatorsBased on ROC Analysis

ConstructNumber of Statements

Cut-Off Chosen

Number Statements

Tested FLAGS II

Persistence 4 2 4Negative Consequences 14 3 14Preoccupation: Obsession 2 2 5Impaired Control: Begin 3 2 4Risky Practices: Later 6 2 7

Risky Practices: Earlier 6 3 6

Impaired Control: Continue 5 3 5

Preoccupation: Desire 4 3 4

Risky Cognitions: Motives 4 2 4

Risky Cognitions: Beliefs 5 3 6

Total 53 5916© Focal Research

Construct and Statements

Co

mp

. 1

Load

ing

Fre

qu

en

cy Hie

r.

Ran

k

Co

rr

wit

h

PG

SI

I often spend more money gambling than I intended.0.86 24.3% 62 0.63

Even when I intend to spend a few dollars gambling,

I often end up spending much more.

0.85 25.9% 55 0.63

I sometimes gamble with money that I can’t really

afford to lose.0.78 21.4% 65 0.70

Once I have started gambling on the slots I find it

very hard to stop..74 24.1% 59 0.64

I often spend more time gambling than I intend to..71 24.1% 56 0.60

Construct and Statements

Co

mp

. 2

Load

ing

Fre

qu

en

cy Hie

r.

Ran

k

Co

rr

wit

h

PG

SI

I have tried to cut back on my slots play with little

success.0.87 9.1% 104 0.68

I have tried unsuccessfully to stop or reduce my

gambling on the slots.0.85 8.0% 109 0.62

There have been times I have gambled despite my

desire not to.0.74 15.0% 76 0.61

Imp

aire

d C

on

tro

l Co

nti

nu

eIm

pai

red

Co

ntr

ol B

egin

17© Focal Research

Discriminant Validity among Reflective Constructs: Gefen and Straub (2005) compared the correlations between the individual items and the PLS calculated

construct scores as listed with the individual items. Min 0.10 diff needed.

Constructs

Statements

Pre

occ

up

atio

n

De

sire

Pre

occ

up

atio

n:

Ob

sess

ed

Imp

aire

d C

on

tro

l:

Co

nti

nu

e

Imp

aire

d C

on

tro

l:

Be

gin

Pe

rsis

ten

ce

Ris

ky C

ogn

itio

ns:

Be

liefs

Ris

ky C

ogn

itio

ns:

Mo

tive

s

Ris

ky P

ract

ice

s:

Earl

ier

Ris

ky P

ract

ice

s:

Late

r

Ne

gati

ve

Co

nse

qu

en

ces

IC-cont1 0.57 0.41 0.87 0.49 0.55 0.29 0.44 0.69 0.53 0.49

IC-cont2 0.50 0.46 0.83 0.55 0.57 0.33 0.46 0.63 0.52 0.50

IC-cont3 0.53 0.43 0.79 0.53 0.51 0.30 0.45 0.58 0.42 0.47

IC-cont4 0.59 0.42 0.89 0.52 0.58 0.31 0.45 0.68 0.54 0.46

IC-cont5 0.49 0.46 0.81 0.49 0.60 0.29 0.43 0.62 0.51 0.54

IC-begin1 0.32 0.48 0.55 0.89 0.66 0.36 0.44 0.49 0.59 0.61

IC-begin2 0.38 0.46 0.54 0.91 0.65 0.35 0.38 0.55 0.62 0.67

IC-begin3 0.36 0.35 0.51 0.80 0.58 0.33 0.42 0.51 0.48 0.54

Persist1 0.32 0.54 0.47 0.61 0.81 0.44 0.56 0.45 0.62 0.66

Persist2 0.37 0.48 0.56 0.58 0.79 0.36 0.47 0.53 0.63 0.65

Persist3 0.39 0.52 0.58 0.57 0.84 0.37 0.49 0.59 0.63 0.64

Persist4 0.39 0.54 0.61 0.66 0.89 0.40 0.56 0.58 0.63 0.7418© Focal Research

Square Root of AVE and Inter-construct Correlations to Test for Divergent Validity (Compeau et al. 1999). Adequate discriminant validity is indicated if the square root

of the construct’s AVE is greater than its correlations with the other constructs.

Erro

ne

ou

s C

ogn

itio

ns:

Be

liefs

Pre

occ

up

atio

n D

esi

re

Ris

ky P

ract

ice

s: E

arlie

r

Imp

aire

d C

on

tro

l:

Co

nti

nu

e

Erro

ne

ou

s C

ogn

itio

ns:

Mo

tive

s

Pre

occ

up

atio

n: O

bse

sse

d

Imp

aire

d C

on

tro

l: B

egi

n

Ris

ky P

ract

ice

s: L

ate

r

Ne

gati

ve C

on

seq

ue

nce

s

Pe

rsis

ten

ce

Erroneous Cognitions:

Beliefsn.a.

Preoccupation Desire 0.40 0.80

Risky Practices: Earlier 0.43 0.54 n.a.

Impaired Control: Continue 0.36 0.64 0.76 0.84

Erroneous Cognitions:

Motives0.49 0.39 0.58 0.53 n.a.

Preoccupation: Obsessed 0.50 0.44 0.50 0.52 0.55 0.83

Impaired Control: Begin 0.40 0.41 0.59 0.61 0.48 0.50 0.87

Risky Practices: Later 0.45 0.48 0.67 0.60 0.59 0.61 0.66 n.a.

Negative Consequences 0.44 0.40 0.61 0.59 0.62 0.56 0.70 0.74 n.a.

Persistence 0.47 0.44 0.65 0.67 0.63 0.63 0.73 0.75 0.81 0.8319© Focal Research

Measures of Internal ConsistencyAverage

Variance

Extracted (AVE)

Composite

Reliability

Average

Communality

Erroneous Cognitions: Beliefs

(Formative)na na 0.31

Erroneous Cognitions: Motives

(Formative)na na 0.49

Preoccupation Desire (Reflective) 0.64 0.88 0.64

Preoccupation: Obsessed (Reflective) 0.68 0.81 0.68

Impaired Control: Begin (Reflective) 0.75 0.90 0.75

Impaired Control: Continue (Reflective) 0.70 0.92 0.70

Risky Practices: Earlier (Formative) na na 0.40

Risky Practices: Later (Formative) na na 0.51

Negative Consequences (Formative) na na 0.41

Persistence (Reflective) 0.69 0.90 0.69

Form

ative C

on

structs

20© Focal Research

Sample Formative Constructs: Risky Practices Early and Later

Risky Practices Early Statement FrequencyHeirarchy Rank R PGSI VIF

I sometimes exceed the amount of money I intended to spend in order to win back money I have lost. 29.7% 40 0.71 1.52

When gambling on the slots I usually use my bank or debit card to get more money so I can keep playing. 19.8% 53 0.53 1.58

I play max bet if I’m on a winning streak. 29.7% 6 0.28 1.15

If I win big I am likely to put the money back into a machine and keep playing. 17.6% 51 0.40 1.34

When gambling on a slot machine I usually play as fast as I can. 15.0% 39 0.36 1.17

I have sometimes gambled for more than six hours straight when I was playing the slots. 34.8% 11 0.41 1.24

Risky Practices Later Statement FrequencyHeirarchy Rank R PGSI VIF

After losing more money than I wanted on the slots I usually try to win it back by playing again either later that day or on another day. 14.7% 81 0.56 1.63

When gambling on the slots I usually use my credit card to get more money so I can keep playing. 10.7% 84 0.42 1.37When I gamble with friends or family I sometimes stay and continue to play after they have stopped or left. 8.0% 103 0.41 1.64

I have sometimes borrowed money from others so I could go and gamble on the slots. 4.3% 111 0.34 1.75I have borrowed money from other people at the casino in order to continue gambling. 4.0% 117 0.35 1.84I have left the casino to get more money so I can come back and keep on gambling. 7.8% 97 0.40 1.60 21

© Focal Research

ROC for Negative Consequences using PGSI Score 8+ for State Variable Value

1 - Specificity

1.00.80.60.40.20.0

Sen

siti

vity

1.0

0.8

0.6

0.4

0.2

0.0

22© Focal Research

Coordinates of the Curve for Consequences and PGSI 8+

Positive if Greater Than or Equal To

Sensitivity (Proportion)

1 – Specificity (Proportion)

0 1.000 1.000

1 0.952 0.204

2 0.952 0.088

3 0.905 0.051

4 0.667 0.023

5 0.667 0.017

6 0.476 0.011

7 0.429 0.006

8 0.429 0.003

9 0.333 0.000

10 0.238 0.000

11 0.143 0.000

12 0.048 0.000

23© Focal Research

FLAGS (EGM)Preliminary Results of Phase IV: development of the FocaL Adult

Gambling Screen

Tony Schellinck, Focal Research and Dalhousie University

Tracy Schrans, President of Focal Research

Michael Bliemel , Dalhousie University

Heather Schellinck, Dalhousie University

Research funded by Ontario Problem Gambling Research Centre – still under review

24© Focal Research