NORTHERN IRELAND DRINK DRIVING RECONVICTION ANALYSIS OF THOSE REFERRED ONTO A COURSE FOR DRINK DRIVING OFFENDERS (2001-2009) - September 2010

 
CONTINUE READING
NORTHERN IRELAND DRINK DRIVING RECONVICTION
ANALYSIS OF THOSE REFERRED ONTO A COURSE FOR
       DRINK DRIVING OFFENDERS (2001-2009)
  September 2010
Table of Contents

                                               Page
Executive Summary                               2

Background                                      4

Previous Research                               5

Methodology                                     7

Results                                         10

Appendices
     Appendix A: Drink Driving Offence Codes    17
     Appendix B: Technical Notes                19

                                                1
Executive Summary

In Northern Ireland between 1 January 2001 and 30 June 2006, 4,637 of those referred
by a magistrate onto the ‘Course for Drink Drive Offenders’ were matched successfully
with conviction data from the Causeway Criminal Record Viewer to enable a
reconviction analysis to be carried out. Of those successfully matched, 1,452 (31%)
completed the course, 69% did not complete a course.

The offenders in the database were all convicted before 1 July 2006, providing a
minimum period of three years between the initial conviction date for the criterion
offence and 30 June 2009 (the cut off period).

Key Findings

   •   A greater proportion of females (35%) than males (31%) who were referred onto
       a course for drink drive offenders opted to attend and complete a course.

   •   Among offenders who had completed the course, 7.0% were reconvicted of a
       subsequent drink drive offence compared to 11.4% who did not complete a
       course. In other words, the likelihood of reconviction was 1.6 times higher if the
       offender didn’t complete a course.

   •   Whilst there was no gender difference in the likelihood of being reconvicted for all
       those who completed a course, males who did not complete a course were found
       to be 1.7 times more likely to reoffend than females who did not complete a
       course.

   •   Up to two years (24 months) after their original conviction, 1.9% of offenders who
       had completed a course had been reconvicted of a subsequent drink/drive
       offence, compared with 6.2% of those who did not complete a course. This
       represents a reconviction rate that is 3.3 times higher for offenders who did not
       complete a course compared with those who did.

                                                                                   2
•   Up to three years (36 months) after their original conviction, 3.4% of offenders
    who had completed a course had been reconvicted of a subsequent drink/drive
    offence, compared with 8.0% of those who did not complete a course. This
    equates to a reconviction rate that is 2.3 times higher for offenders who did not
    complete a course compared with those who did.

•   The effectiveness of the courses in NI appears to broadly reflect the experience
    in GB where it was similarly observed that the largest impact occurs in the first
    two years following initial conviction.

                                                                             3
Background

The Road Traffic Offenders (Northern Ireland) Order 1996 provided the opportunity to
trial the use of rehabilitation schemes for drink drivers. These schemes were developed
as an educational intervention aimed at reducing the likelihood of re-offending by
individuals who have been convicted of a drink drive offence. Courts can be given
powers to reduce the period of disqualification (up to a maximum of 25%) if the offender
satisfactorily completes a rehabilitation course approved by the Department of
Environment Northern Ireland (DoE NI).

The Northern Ireland (NI) Courses for Drink Drive Offenders (CDDO) pilot scheme
formally commenced on 1 April 1998, in two designated Magistrate Court areas - Belfast
and Newtownabbey. Although Probation Board for Northern Ireland (PBNI) was the sole
course facilitator, there was also a multi- agency element in the courses, with input from
the Police Service of Northern Ireland (PSNI), DoE Road Safety Education Officers,
Campaign Against Drink Drivers and Road Trauma Support. Legislation was introduced
on 1 January 2001, to extend the existing pilot scheme until 31 December 2005, with
the NI Court Service extending the authority to refer relevant offenders to all Magistrate
courts in Northern Ireland. This scheme was made permanent following the successful
evaluation in 2005.

From the outset, the evaluation of the scheme has been core to its implementation, with
the aim being;
“to allow the use of such courses, as a sentence mechanism, to be evaluated in terms
of their effectiveness in reducing drinking and driving”.

This latest evaluation was commissioned by DoE NI, by way of the CDDO working
group.   The CDDO working group is chaired by an official from DOE Road Safety
Division and has included representatives from the Department of Justice, the Public
Prosecution Service, the Probation Board, the Driver and Vehicle Licensing Agency,
Road Safety Division and Central Statistics and Research Branch. The purpose of this
type of evaluation is to add to the qualitative information already gathered and to inform
the CDDO working group of the effectiveness of the scheme by studying reconviction
rates.

                                                                                  4
Previous Research

Northern Ireland
A reconviction study was carried out in Northern Ireland in 2005 to determine whether
the CDDO had any effect on the relative reconviction rates of those who were referred
onto and completed a course compared to those referred onto but who did not complete
a course. Between 1 April 1998 and 31 December 2000, 370 people were referred by a
magistrate onto a course. Thirty-two percent completed a course and 68% did not
complete a course.

Up to two years (24 months) after their original conviction, 0.9% of offenders who had
completed a course had been reconvicted of a subsequent drink/drive offence,
compared with 5.7% of those who did not complete a course. The reconviction rate of
those who did not complete a course was 6.3 times the rate of those who did complete
a course.

Up to three years (36 months) after their original conviction, 2.6% of offenders who had
completed a course had been reconvicted of a subsequent drink/drive offence,
compared with 9.8% of those who did not complete a course. The reconviction rate of
those who did not complete a course was 3.8 times the rate of those who did complete
a course.

Great Britain
Department for Transport (DfT) (formerly DETR) commissioned the Transport Research
Laboratory (TRL) to undertake research into calculating reconviction rates for this
scheme in Great Britain (GB). Results of this were published in 1999 and can be found
in TRL report 426, ‘Drink/driver rehabilitation courses in England and Wales’. In short,
TRL studied 3 groups of people who were convicted between 1993 and July 1996:
   •   Those convicted and who attended a course (at an experimental court).
   •   Those convicted and not attending (at an experimental court).
   •   Those convicted at a control court (where the scheme was not available).

These groups where ‘followed’ until December 1998, allowing TRL to base
reconviction rates on at least 29 months after, and in some cases 5 years after
the first offence.

                                                                                  5
The purpose of the control group was to test whether any observed difference in
reconviction rates between those attending and not attending courses could be
due to ‘subject selection bias’, i.e., the possibility that offenders predisposed
against reoffending were more likely to be offered, accept and complete training
than offenders predisposed to reoffend. In the event, whilst no real difference in
reconviction rates was found between all offenders convicted in the experimental
courts and those in the control courts, further more detailed analysis concluded
that subject selection bias only accounted for a small element of the differences
in reconviction rates reported below.

Twenty-four months after their original conviction, only 2% of offenders who had
attended courses had been convicted of a subsequent drink/drive offence, compared
with 6% of those who had not in the experimental courts. The reoffending rate of those
who did not attend a course was 3 times the rate of those who did.

Thirty-six months after their original conviction, only 3.4% of offenders who had
attended courses had been convicted of a subsequent drink/drive offence, compared
with 9.6% of those who had not in the experimental courts. The reoffending rate of
those who did not attend a course was 2.8 times the rate of those who did, still very
similar to the twenty-four month ratio.

TRL carried out a further study in GB, the results of which were published in 2003 and
can be found in TRL report 574, ‘Reconvictions for drink/drive course attenders: a six
year follow up’.    This report examines the reconvictions of the original sample of
offenders tracked in the experimental courts, in report 426, only over a longer period of
time providing at least 72 months of data for over 80 percent of the sample.

TRL’s 6-year follow-up study found that after 72 months, nearly 2.5 times more non-
course attenders than course attenders had reoffended (17.9% of non-course attenders
compared with 7.6% of course attenders). This ratio of rate of reconvictions for non-
course attenders to the rate of course attenders was found to be similar at 3, 4, 5 and 6
years after conviction.

                                                                                     6
Methodology

In the original NI research, court data was not available for those people who were not
offered the scheme from 1 April 1998 to 30 December 2000, and therefore it was not
possible to have a control group. Whilst the GB study could not be replicated in its
entirety, DRD statisticians followed its basic structure. The new research carried out in
2010 was set-up primarily to replicate the original NI research, therefore the same
methodology has been applied, i.e, without a ‘no course offered’ control group.

In GB it was concluded that ‘subject selection’ was not a significant source of bias. It is
not unreasonable to assume that this conclusion would similarly apply in NI. However,
it would be preferable that this actually be tested in the local context. It is anticipated
that, with increased resources and availability of court data, further analysis will be
carried out in this area which will allow this possibility to be properly investigated in
future.

The initial evaluation in 2005 sought to inform the CDDO Working Group of the
effectiveness of the course, with the reconviction analysis supplementing the attendees’
pre and post course knowledge of alcohol and attitudes to drinking and driving. The
analysis carried out in 2010 focuses solely on reconviction rates.

For the purposes of the original first wave NI evaluation, the groups of drivers followed
were those who had been convicted of a relevant drink driving offence between 1 April
1998 and 31 December 2000 (the pilot period) and were referred by a magistrate to
attend one of the drink drive courses. The pilot period was chosen for the study so that
reconviction rates could be examined over the maximum time period possible at that
time. The drivers in this group were then tracked up to 30 June 2004. This allowed an
investigation of the relative reconviction rates of drink/drive offenders who, once
referred, had attended a course compared to those who had not.

This new analysis is based on a much larger population of those convicted of a relevant
drink drive offence during the period 1 January 2001 and 30 June 2006 and referred by
a magistrate to attend a drink drive course. The referrals were tracked up to 30 June

                                                                                   7
2009. These time periods were selected to ensure that each person referred had a
minimum of a three year period in which they could be reconvicted.

Data

The original evaluation used two main sources of data:

   i)     The PBNI referral database (CDDO) - which holds details of all those persons
          convicted of a relevant drink drive offence (see Annex A) and who were
          referred onto a course by a magistrate.
   ii)    Causeway Criminal Record Viewer (CRV). - This database records
          information of all offences and disposals in all NI courts.

These data sources were utilised again for the 2010 analysis.

Between 1 January 2001 and 30 June 2006, PBNI had provided DRD with details of
5,967 people who were referred by a magistrate onto a CDDO course. The CDDO
referral database was used as the ‘population’ and the individual criminal records of
those sentenced between 1 January 2001 and 30 June 2006 were accessed via an
automated match with the CRV by DRD and Department of Justice (DoJ) (formerly
Northern Ireland Office) statisticians. A ‘unique’ identifier for each referral was populated
initially using the surname, forename and date of birth, as given on the CDDO referral
database and the CRV.        A further match with ’date of sentence’ was carried out.
Despite some data issues, particularly with regard to incomplete and missing
name/address information, a total of 4,637 individual referrals were matched
successfully providing an overall match rate of 78%.      Referrals that were not matched
have been disregarded from this analysis. There is no reason to believe non-matches
would be anything other than random and their exclusion therefore should not introduce
any systematic bias into the study.

Because of the exclusions, it is important to be aware that any references to resultant
matched number of referrals do not represent the total number of all those referred onto
CDDO courses.

                                                                                    8
A reconviction database was created by DRD and DoJ with the following additional
information being added:
   -   Date of any further reconvictions for relevant drink driving offence
   -   Indication of whether individual has been reconvicted of relevant offence
   -   Number of reconvictions
   -   End date of 30 June 2009 (research cut off period)
   -   Indication of whether CDDO course had been completed
   -   Reconviction period between primary and subsequent offence

Responsibility for the maintenance of the referral database during this period resided
with Central Statistics and Research Branch, DRD. Statisticians from DRD and DoJ
agreed the search and match methodology outlined previously, and the recording of
additional information on the database from the CRV. Access to the CRV was controlled
by DoJ statisticians. Permissions to match and access the new database was managed
through an access agreement between DRD and DoJ. Once the referral and
reconviction database was complete with all necessary information, DoJ transferred it to
Central Statistics and Research Branch DRD so that the necessary analysis and
evaluation could be carried out. The database is securely stored within the branch with
access restricted to only those Road Safety statisticians who need to undertake any
analysis. The DRD and DoJ staff involved with the study are on secondment from the
Northern Ireland Statistics and Research Agency. This ensures that the data
collection, matching, validation, analysis and evaluation were carried out by
independent statisticians bound by the Code of Practice for Official Statistics.

                                                                                   9
Results

Note that not all percentages presented in this section may sum to 100 due to rounding.

Study Group - Profile

Out of the 4,637 individuals matched and which formed the study population, 3,185
(69%) opted not to attend a course and 1,452 (31%) opted to attend and actually
complete a course. See Figure 1.

Figure 1: Number of people completing the course by gender

                    3500
                                                                                 3185
                    3000                                          2742
 Number of People

                    2500

                    2000
                                        1452
                    1500
                           1217
                    1000
                                                                           443
                    500
                                  235
                      0
                                  Yes                                      No

                                               Completed Course
                                        Male    Female   Male and Female

Whilst the overwhelming majority of course referrals during the study period were males
(85%), a greater proportion of females (35%) than males (31%) actually completed the
course.

The most common period of disqualification for all who were referred onto a course was
12 months, with 79% of offenders being disqualified from driving for this length of time.
Thirteen percent were disqualified for more than 12 months and less than or equal to 24
months, 6% were disqualified for more than 24 months, but less than or equal to 36
months and 1% were disqualified for over 36 months.

The offenders in the database were all convicted before 1 July 2006, hence the
minimum time between conviction for the criterion offence and 30 June 2009 (research
cut off period) was 36 months with the maximum being 102 months (or 8.5 years). See
Figure 2 and Table 1 overleaf.

                                                                                        10
Figure 2: Number of months between criterion conviction offence and 30 June 2009

                    1800
                                                                                    1674
                    1600
                    1400
 Number of People

                    1200
                               1027          1021
                                                                    915
                    1000
                     800
                     600

                     400
                     200
                       0
                               36-48         49-60              61-72                73+
                                                 Number of Months

Table 1 – Time Profile between criterion offence and 30 June 2009
                      Months           Number                 %
                      36-48              1027                 22.1
                      49-60              1021                 22.0
                      61-72               915                 19.7
                      73+                1674                 36.1
                      Total              4637                  100
Note: Due to rounding percentages may not sum to 100.

In order to ensure that any observed difference in reconviction rates, between those
who completed courses and those who did not, was not simply due to a differential
length of time available in which to reoffend, the time of first conviction profile of each
group was also compared (See Table 2).

Table 2 – Percentage of first convictions by year of conviction and course completion
status
Year of First Conviction                Individuals Completing a             Individuals Not
                                               Course (%)                 Completing a Course (%)
2001                                               10                               12
2002                                               13                               15
2003                                               21                               19
2004                                               21                               21
2005                                               23                               21
2006                                               12                               11
All First Convictions                             1452                             3185

                                                                                                    11
Whilst there are small differences between profiles (typically 2 percentage points or
less), statistical testing shows that there is no overall systematic difference between the
groups. This provides additional evidence that any difference in the overall reconviction
rates can be attributed to the influence of the course.

Reconviction Rates

Overall, between 1 January 2001 and 30 June 2009, 466 people out of the 4,637 study
population were reconvicted of a further drink/drive offence – 10%.

Among offenders who had completed the course, 102 out of 1,452 were reconvicted -
7.0%. The proportion of males being reconvicted after completing a course was the
same as that for females.

Among offenders who were referred, but who did not complete a course, 364 out of
3,185 were reconvicted - 11.4%. The percentage of males who did not complete a
course and were reconvicted (12%) was 1.7 times greater than the percentage of
females who did not complete a course and were reconvicted (7%).

It would seem reasonable to conclude, therefore, that whilst there is no gender
difference in the reconviction rate of males and females who have completed a course,
non-completion would appear to have a bigger impact on males in terms of increased
likelihood of reoffending.

Over the entire study period, offenders who did not complete a course were 1.6 times
more likely to be reconvicted for a subsequent drink/drive offence, compared with
offenders who had completed a course.

For the 466 persons in the study group who were eventually reconvicted, Figure 3
graphically illustrates the time between the primary and subsequent offence for
completers versus non-completers while Table 3 provides further detail in terms of
numbers and percentages.         It should be noted that, apart from in the 60-72 month
category, the apparent small differences between completers and non-completers in the
proportions reconvicting after the first 2 years are not statistically significant.

                                                                                      12
Figure 3: Time between primary and subsequent offence by course completion status

  3.0%

  2.5%

  2.0%

  1.5%

  1.0%

  0.5%

  0.0%
         0-6 months >6 months >12 months >24 months >36 months >48 months >60 months >72 months
Up to 2 years (24 months) - the reconviction rate of those who did not complete a
course was 3.3 times the rate of those who did complete a course.

Up to 3 years (36 months) - the reconviction rate of those who did not complete a
course was 2.3 times the rate of those who did complete a course.

The reconviction rates for up to four years (48 months) were 4.9% and 9.5% for course
completers and non-completers respectively while up to five years (60 months) they
were 5.8% and 10.5%. For both time periods, the likelihood of reoffending is just under
2 times greater for those who complete the course. However, as the four and five year
rates are based on an increasingly smaller subset of the total study group, they are not
as robust as the estimates up to 3 years which comprised the total group.

An alternative way of looking at this is presented in Figure 4 overleaf which illustrates
that a larger proportion of course completers remained reconviction free across all time
periods. It also shows the markedly increasing rate of reconviction within the first two
years for non-completers compared to those who completed a course. The reconviction
rate then appears reasonably consistent between the groups for the remainder of the
study period. This suggests that the effectiveness of the course peaks in the first two
years following conviction, a finding also highlighted in Figure 3 above.

Figure 4: Proportion WITHOUT further drink drive convictions by course completion
status
                                    1.00

                                    0.98
 Proportion without reconvictions

                                    0.96

                                    0.94

                                    0.92

                                    0.90

                                    0.88

                                    0.86

                                    0.84

                                    0.82
                                           0-6 months   >6122436486072 months
                                                        months    months        months      months         months      months
                                                                           Time since criterion offence
                                                                 Course completed                    Course not completed

                                                                                                                                             14
Of the 1,452 people who did complete a course, 1,350 (93.0%) had no further
convictions during the study period, 91 had one further reconviction (6.3%) and 11 had
2-3 more convictions (0.8%).

Of the 3,185 people who were referred but who did not complete a course during the
study period:
2,821 (88.6%) had no further convictions;
319 (10.0%) had one further reconviction;
32 (1.0%) had two further reconvictions;
8 (0.3%) had three further reconvictions; and
5 (0.2%) had more than three further reconvictions.

Comparison with NI First Wave Results and GB

This research reports the reconviction rate of those completing a course to be 3.3 times
that of those not completing a course, for the period up to 2 years after their criterion
conviction. This ratio falls to 2.3 for the period up to 3 years after conviction. The
original research carried out in NI reported the reconviction rate of those completing a
course to be 6.3 times that of those not completing a course for the period up to 2 years
after their criterion conviction. This ratio fell to 3.8 for the period up to 3 years after
criterion conviction. These ratios for up to 2 years and up to 3 years are evidently lower
in this new research. However, it should be borne in mind, that the significantly larger
sample size in the current research will yield much more robust results.

The research carried out in GB in 1999 by TRL found that 2% of course attenders and
6% of non-attenders had been reconvicted over the period up to 2 years after criterion
conviction. For the period up to 3 years after criterion conviction TRL found that 3.4% of
course attenders and 9.6% of non-course attenders had been reconvicted.               This
translates into a ratio of 3 for the period up to 2 years after criterion conviction and a
ratio of 2.8 for the period up to 3 years.

Apart from perhaps the year 3 reconviction rate for non-attenders, these results are
strikingly similar to the current NI 2 and 3-year rates for attenders (1.9% and 3.4%) and
non-attenders (6.2% and 8.0%) noted above.

                                                                                 15
TRL’s 6-year follow-up study found that after 72 months, nearly 2.5 times more non-
course attenders than course attenders had reoffended (17.9% of non-course attenders
compared with 7.6% of course attenders). This ratio of rate of reconvictions for non-
course attenders to the rate of course attenders was found to be similar at 3, 4, 5 and 6
years after conviction. It fell from just under 4 in the period up to 2 years after criterion
conviction to just under 3 in the period up to 3 years after criterion conviction with only a
relatively smaller decline to 2.5 after 6 years. Whilst these latest findings might suggest
a stronger impact of the courses in GB, it should be noted that the GB results are based
on a much larger study group and hence will be more robust than the equivalent NI
findings.

The trend identified in NI is broadly comparable to that found in the GB research above
i.e., that the greatest impact of the course occurs in the first 2 years followed by a
levelling off in terms of course effectiveness. TRL, in 2003, highlighted that the impact of
the courses appears to be more effective during the driver disqualification period.
Whatever the reason, the positive impact noted in the original NI research has been
reaffirmed in this latest study.

                                                                                   16
ANNEX A
Drink Driving Offence Codes

N.I.OFFENCE CODEOFFENCE DESCRIPTION
50 08 044        Failure to provide specimen of blood/urine - major
50 08 046        Driving with excess alcohol
50 08 048        In charge with excess alcohol
50 08 049        Driving with excess alcohol – blood /urine
50 08 052        Failing to provide specimen of breath - driving
50 08 053        Failing to provide specimen of breath – attempting to drive
50 08 123        Driving while unfit by drink or drugs
50 08 124        Attempting to drive while unfit by drink or drugs
50 08 125        In charge of a vehicle while unfit by drink or drugs
50 08 126        Driving with excess alcohol in breath
50 08 127        Attempting to drive with excess alcohol in breath
50 08 128        In charge of a vehicle with excess alcohol in breath
50 08 129        Driving with excess alcohol in blood
50 08 130        Attempting to drive with excess alcohol in blood
50 08 131        In charge of a vehicle with excess alcohol in blood
50 08 132        Driving with excess alcohol in urine
50 08 134        In charge of a vehicle with excess alcohol in urine

50 08 136        Failing to provide a specimen of breath for analysis
50 08 137        Failing to produce a specimen of blood or urine for analysis
50 08 138        Failing to produce a specimen when driving unfit
50 08 139        Failing to provide a specimen when attempting to drive unfit
50 08 140        Failing to provide a specimen when in charge unfit
50 08 141        Failing to provide a specimen when driving with excess alcohol
50 08 142        Failing to provide a specimen when attempting to drive with
                 excess alcohol
50 08 143        Failing to provide a specimen when in charge with excess
                 alcohol

50 08 204        Causing death by driving carelessly when unfit
50 08 206        Cause GBI by driving carelessly when unfit
50 08 207        Cause GBI by inconsiderate driving when unfit

                                                                               17
N.I.OFFENCE CODEOFFENCE DESCRIPTION
50 08 208      Causing death by driving carelessly with excess alcohol
50 08 209      Causing death by inconsiderate driving with excess alcohol
50 08 210      Cause GBI by driving carelessly with excess alcohol
50 08 211      Cause GBI by inconsiderate driving with excess alcohol

50 08 212      Cause death by driving carelessly and failing to provide
               specimen
50 08 214      Cause GBI by driving carelessly and failing to provide
               specimen

                                                                          18
ANNEX B
Technical Notes

Data Matching

Due to incomplete data and missing variable information such as surnames, date of
births, conflicting gender etc. a number of assumptions have been built into the data set
which are detailed below:

1) Initial match of CDDO referral database with CRV was on a joined variable containing
   both DOB and Surname. Initially this match only identified over 3,000 cases. Further
   matches were then carried out adding in forename and CRV data of conviction to
   establish a final match of 4,637 cases.
2) A match was carried out to calculate the date of referral from the CDDO database
   against the conviction date on the CRV database. Based on the results a number of
   convictions that occurred prior to the CDDO referral date were removed from the
   database. Where there was a negative result all differences were inspected prior to
   any changes being made to the file.
3) During analysis there were @ 50 cases where gender conflicted between the two
   datasets. An individual check against forename was carried out and a gender marker
   set on this basis.
4) There were approx 90 cases were CRV conviction court date conflicted with the
   CDDO court date, each of these cases were individually inspected and dates
   amended as appropriate against other date information where available.            For
   approximately 10 cases a date was selected by the statistician as it was not apparent
   which date was incorrect.
5) In 4 cases the match with the CDDO database and the CRV populated dates that
   were vastly apart. For the purposes of this analysis an additional row was added to
   the dataset for the original referral date and then the latter date was treated as a
   reconviction.
6) In 32 cases the number of months disqualified were missing from the CDDO
   database, these missing values were populated from the CRV length of time
   disqualified variable.

                                                                              19
7) In the CDDO database there were a number of duplicate cases which were removed
   from the database and not included in the analysis. These were checked to ensure
   that they were not re-referrals.

Significance Testing

The results presented in this report cover a discreet study period and could be regarded
as a sample of all persons who have been referred onto a course for drink drive
offenders. As such the results will be subject to a degree of sampling error.

However, any statements in this report regarding differences between groups, such as
those who have completed a course and those who have not, are statistically significant
at the 95% significance level. This significance level is the usual threshold for judging
whether differences between groups might have arisen by chance. It means that there is
only a 5%, or one in twenty, probability that an observed difference might have arisen by
chance factors alone.

                                                                                20
You can also read