(Ebook PDF) Foundations of Statistics (A Pearson Original) 4th Edition Online Ebook Textbook Full Chapter PDF
(Ebook PDF) Foundations of Statistics (A Pearson Original) 4th Edition Online Ebook Textbook Full Chapter PDF
OR CLICK LINK
http://ebooksecure.com/product/ebook-pdf-
foundations-of-statistics-a-pearson-original-4th-
edition/
Download More ebooks [PDF]. Format PDF ebook download PDF KINDLE.
Full download test bank at ebooksecure.com
More products digital (pdf, epub, mobi) instant
download maybe you interests ...
https://ebooksecure.com/download/progress-in-heterocyclic-
chemistry-ebook-pdf/
http://ebooksecure.com/product/ebook-pdf-translational-medicine-
in-cns-drug-development-volume-29/
https://ebooksecure.com/download/netter-atlas-of-human-anatomy-
classic-regional-approach-8e-mar-29-2022_0323793738_elsevier-not-
true-pdf-ebook-pdf/
https://ebooksecure.com/download/cardiology-an-integrated-
approach-human-organ-systems-dec-29-2017_007179154x_mcgraw-hill-
ebook-pdf/
(Original PDF) Australasian Business Statistics, 4th
Edition
http://ebooksecure.com/product/original-pdf-australasian-
business-statistics-4th-edition/
http://ebooksecure.com/product/ebook-pdf-statistics-for-business-
pearson-new-international/
http://ebooksecure.com/product/ebook-pdf-foundations-of-
economics-4th-edition/
http://ebooksecure.com/product/original-pdf-foundations-of-
taxation-law-2018-10th-edition/
http://ebooksecure.com/product/ebook-pdf-the-little-pearson-
handbook-4th-edition/
Foundations of Statistics
4th Edition
Dr Glenda Francis
Dr Andi Garing
Module 1 Critical Thinking in Statistics
MODULE1
Introduction to Critical Thinking in Statistics
In both your personal and your professional life you'll often come into contact with statistics.
Open a newspaper, surf the internet or turn on the radio or the television and you're going to
fmd examples - anything from an advertising campaign trying to convince you to use a
particular anti-dandruff shampoo, to a report on the strength of opposition to some
government policy. Did you know that Labor voters prefer to watch soapies, while Liberal
voters prefer cop shows? Can this be true? - well according to a report I heard on the radio, it
is - but should I believe them? Some statistics reported in the media defmitely come into the
category of'lies, damned lies and statistics'but some represent well conducted studies with
important consequences. How can we tell which is which? One of the aims of this text is to
convince you of the value of good statistical studies while proofmg you against being misled
with statistics.
Throughout this text you'll meet a lot of statistical'jargon'. The language in statistics is very
precise, and terms you thou迦 you understood, like'population'and'significant', are used in
rather different ways in statistics to how they are used in everyday English. So in some ways
learning about statistics is like learning another dialect and your language skills will be far
more important than your mathematical skills.
1 . 2 Detecting Bias
So what distinguishes the good statistical studies from the bad? Let's start by looking at a
research question of our own, and explore how we might go about answering it. In 2000, 60%
of Australians reported having access to a computer at home. Has this percentage increased?
How would we go about answering this question? We could start by asking some people if
they had access to a computer at home. Suppose we put this question to a group of people and
recorded their responses. Rather than trying to analyse the data by hand, we'll use a statistical
package SPSS2, an IBM company —IBM SPSS Statistics release 23 (2015)3, (SPSS stands for
' Statistics Package for the Social Sciences'.) You'11 find instructions for using SPSS in
Appendix A, here we'll just show you the output and how to interpret it.
IBM Corp. released 2015. IBM SPSS Statistics for Windows, Version 23.0. Armonk, NY: IBM Corp.
Copyright ©2016 Pearson Australia (a division of Pearson Australia Group Pty Ltd) ISBN (9781488613432), Dr. Andi Garing, Foundations of Statistics, 4th Edition
Module 1 Critical Thinking in Statistics
computer at home?
Valid yes
Percent
88.3
I Valid Percent
88.3
l Cumulative
Percent
88.3
no 11.7 100.0
Total 100.0
So in this sample of 103 people, 8~.3% had a computer at home. That's a lot more than the
60% of Australians who reported having a home computer in 2000. Is this enough
information to convince you that the ownership of home computers has increased? This
question raises some important issues.
Copyright © 2016 Pearson Australia (a division of Pearson Australia Group Pty Ltd) ISBN (9781488613432), Dr. Andi Garing, Foundations of Statistics, 4th Edition
Module 1Critical Thinking in Statistics
4 For more information on subjective survey data, see Bertrand, M. & Mullainathan, S. "Do People Mean What
They Say? Implications for Subjective Survey Data" Economics and Social Behavior
5 Loftus, E. & Palmer, J. 1974 "Reconstruction of Automobile Destruction: An Example of the Interaction
Between Language and Memory" Journal of Verbal Learning and Verbal Behaviour 13, 585-589
4
Copyright © 2016 Pearson Australia (a division of Pearson Australia Group Pty Ltd) ISBN (9781488613432), Dr. A ndi Garing, Foundations of Statistics, 4th Edition
Module 1 Critical Thinking in Statistics
Even the order in which questions are asked can have an effect. The response to'How happy
are you with life in general? ' might depend on what other questions the respondents have just
been asked. Previous questions might remind them of some aspect of their life they are
particularly happy (or u呻appy) with, which tends to colour their response to the'general
happiness'question.
Problems with the way measurements are taken or with the way questions are asked can
introduce substantial bias. These effects might be unintentional or deliberate, and they might
be obvious, as in a leading question'Should refugees be subjected to the appalling conditions
in detention centres?'or far more subtle.
~EXERCISE 1.3 (Solution on page 255)
Comment on any obvious bias in the following measurements.
a. A research organisation is conducting a study on healthy eating habits. In this study the
respondents are asked to report on the number of take away meals they ate last week.
b. In a study on healthy lifestyles, a nurse records the weight of each of the participants
each week.
c. A market researcher asks shoppers to try out a new product for cleaning glass surfaces.
After the trial he asks'Is this remarkable product better than your normal cleaner?'
d. A market researcher asks a sample of television viewers: "Do you enjoy life style
programs?"
So what's the overall conclusion here? Whenever you are looking at the results from any
statistical study, you need to think critically. You should always be looking for possible bias.
Here are a few specific things to keep in mind:
Copyright©2016 Pearson Australia (a division of Pearson Australia Group Pty Ltd) ISBN (9781488613432), Dr. A ndi Garing, Foundations of Statistics, 4th Edition
Module 1 Critical Thinking in Statistics
The closure of the abattoir has destroyed the livelihoods of many people m our
community, do you agree that it should be re-opened?
1. Yes
2. No
When they analyse the data they report back to AMIC that 97% of the respondents eat
meat every day, and that 95% agreed that the abattoir should be re-opened. AMIC feels
very encouraged. In a letter to the owners of the abattoir they report that interest in red
meat is very high with 97% of Australians eating red meat at least once a day. They also
report that 95% of people support the re-opening of the abattoir.
List all the possible sources of bias you can see in this study. (I can see at least 8). As an
extension question, and a challenge for you —come up with an equally ludicrous scenario
involving even more bias than this one :-).
Copyright © 2016 Pearson Australia (a division of Pearson Australia Group Pty Ltd) ISBN (9781488613432), Dr. A ndi Garing, Foundations of Statistics, 4th Edition
Module 1 Critical Thinking in Statistics
So if 70 out of 1000 people supported tree planting, the proportion of people who support
more tree planting would be:
p roportwn = 70 = 0.07
1000
To get the percentage, just multiply this value by 100. In this case, the percentage of people in
favour of more tree planting was .07 x 100, that is, 7%. (We don't really require much more
mathematical skill in this text, so if you can calculate percentages your maths skills are fine!)
Proportions and percentages give you the same information, but in reports percentages roll off
the tongue more easily; "The proportion of respondents who support the proposal is only .07"
doesn't sound as good as "Only 7% of respondents support the proposal".
Let's look at another couple of examples where percentages are more useful than just the
number of cases. Consider the following excerpt from an article titled "Everyone enjoying
,,6
boom times: Costello
"More people were in work than ever before in Australian history and there were two
million more jobs than a decade ago, Mr Costello told ABC Radio."
Two million more jobs sounds like a lot, but how does this compare to the increase in the
number of adult Australians over that time? We might prefer to know how the proportion of
people in full time employment has changed over that period. (We might also have other
questions about those 2 million jobs: were they all full time jobs or were some of them part
time?)
Here's another example, this one comes from the intemet7, and looks at the number of
industrial accidents reported (I can't tell you what country the figures refer to - there was no
indication on the website of the source of these figures!) In 2005 there were 1300 industrial
accidents reported in the ' construction'industry but only 456 accidents reported in the
'shipbuilding and ship repair'industry. Does this tell us that it's more dangerous to work in
construction than in ship building? Not really- it depends on how many workers there are in
htt ://www.mom. ov.s / ublish/etc/medialib/mom libra /Work lace Safet /files6.Par.67751.File.tm /Numbe
roflndustrialAccidentsbylndustry19962005.pdf accessed 25/2/07
7
Copyright©2016 Pearson Australia (a division of Pearson Australia Group Pty Ltd) ISBN (9781488613432), Dr. Andi Garing, Foundations of Statistics, 4th Edition
Module 1 Critical Thinking in Statistics
each of these industries. If there are far more workers in the construction industry, then we 'd
expect far more industrial accidents in that industry. If we want to make an informed
comparison, we need to know what percentage of workers in each industry are involved in
industrial accidents.
While we often need to know the percentage rather than the number of cases, there are some
situations where it's the actual number that is more important. To take a trivial example, if I
invite 50 people to my birthday party, I want to know how many are going to come, not what
percentage of those I invited are coming. If we go back to our original example about access
to computers at home - when I'm trying to estimate how many copies of the SPSS software I
need to order, I need to know how many students in my subject have access to a computer at
home, not what percentage.
Copyright ©2016 Pearson Australia (a division of Pearson Australia Group Pty Ltd) ISBN (9781488613432), Dr. Andi Garing, Foundations of Statistics, 4th Edition
Module 1 Critical Thinking in Statistics
Copyright©2016 Pearson Australia (a division of Pearson Australia Group Pty Ltd) ISBN (9781488613432), Dr. A ndi Garing, Foundations of Statistics, 4th Edition
Module 1: Critical Thinking in Statistics
1 . 4 Summarising Data
Suppose a reporter wanted to write a report in their magazine about how much time
Australians spend doing housework. They might start by collecting a lot of information from
a carefully selected sample of Australian adults, but how would they present this information?
It certainly wouldn't make much sense to the readers if they just printed all of the raw data
they'd collected without analysing it in some way. We need some tools for summarising
information. Some of these tools, like percentages and the mean (average) you've no doubt
met before, but some, like percentiles and boxplots you may not have seen.
Some Tools
SPSS
Instead of slogging through lots of tedious calculations on a calculator, and drawing graphs by
hand (I was always hopeless at drawing graphs!) we'll get SPSS to do the analysis for us. In
the SPSS data file What_Analysis.sav, we have some relevant data about the lives of
Australian adults - including information on hours of housework - which is based on data
gathered in a government survey. Here's a view of the first part of the data file:
You can see that instead of recording'Male'and 'Female'for sex, number codes have been
used. This is a very common practice. Before we can interpret this information, we need to
know what each of the number codes represents. In this case, sex has been coded female = 0,
male = 1. This was a totally arbitrary choice, we could have used male = 1, female = 2, or
even male = 21, female = 37 ifwe wanted to, but it's best to keep the codes simple!
10
Copyright©2016 Pearson Australia (a division of Pearson Australia Group Pty Ltd) ISBN (9781488613432), Dr. A ndi Garing, Foundations of Statistics, 4th Edition
Module 1: Critical Thinking in Statistics
This coding is recorded as part of the SPSS data file. If we wanted to know what'marital'
represented, and how it was coded, we could just click on 斗圃 in the icon bar and select
'marital'in the list of variables:
Variables
Variable Information:
identification nu... 1.. marital
sex[sex) Label: Marital status
# age in years [age) Type: F5
l.larital status [... Missing Values: 9
, - 『-.一
巨 呜邑丘
So how do we go about summarising the information in this data file? We need a few
statistical tools.
sex
Cumulative
Percent
Valid female 55.4
male 100.0
Total
11
Copyright ©2016 Pearson Australia (a division of Pearson Australia Group Pty Ltd) ISBN (9781488613432), Dr. Andi Garing, Foundations of Statistics, 4th Edition
Module 1: Critical Thinking in Statistics
The frequency table tells us at percentage of people were in each catego . But there are two
columns in the table givin~ ercentages - one headed'Percent'anf ne headed'Valid
Percent'. Which one should w use? In most cases it's the'Valid Percent'that we're interested
in. This tells us that of the 416 people who answered the question, 67.5% said they were
married.
Output from the FREQUENCIES procedure - Pie Charts and Bar Charts
As well as the frequency table for these two variables, we might also want some sort of
graphical representation. So first let's look at Pie Charts.
口 ~心r mom<ld
Figure 1.2. Pie chart for sex. Figure 1.3: Pie chart fo r marital status.
From the pie chart for sex (Figure 1.2) we can see at a glance that there were more females in
the study than males. This gives a quick instant picture of the distribution. But the pie chart
for marital status (Figure 1.3) is not quite so useful. It's a bit too cluttered, and doesn't give a
nice clear picture of what's happening.1 When there are more than two or three categories, pie
charts are not very effective. There is an alternative whic~gives a much clearer picture of the
distribution when there are more than 2 or three categon es. This alternative is a percentage
bar chart. In a bar chart, each category in the variable is represented in a separate column, and
the height of the column represents the percentage of responses in that category. The
percentage bar chart for marital status is displayed in Figure 1.4.
1 Note that the pie chart produced by SPSS version 20 includes the 'missing values'category, and the
percentages shown would not the 'valid percentage'. We would have to edit the pie chart to remove the missing
value segment. Instructions for doing this are included in Appendix A2. In SPSS version 23 the missing value
segment is not included in pie charts.
12
Copyright ©2016 Pearson Australia (a division of Pearson Australi a Group Pty Ltd) ISBN (9781488613432), Dr. AndiGaring, Foundations of Statistics, 4th Edition
Module 1: Critical Thinking in Statistics
Marital status
From the bar chart, it's obvious that
60 the most common marital status was
'married'. (In statistics, this is called
芒只》』
4 the mode.)
。
<Id
c. Of the people in this sample who gave their occupation, what percentage were semi-
professional/managerial?
d. Now use the FREQUENCIES procedure to produce a pie chart for occupation. Do you
think the pie chart or the bar chart gives a clearer picture of the distribution of
occupation?
13
Copyright©2016 Pearson Australia (a division of Pearson Australia Group Pty Ltd) ISBN (9781488613432), Dr. Andi Garing, Foundations of Statistics, 4th Edition
Module 1: Critical Thinking in Statistics
0 1 23 4 56789
Histograms look similar to bar charts, but there are some really important differences. In a bar
chart, each bar represents one category, and each bar is labelled individually. With a
histogram, what we're doing is grouping the data. Note that each of the columns in the
histogram has the same width. So the first column in the histogram tells us how many people
watched from O up to but not including 1 hour of television each day (about 8), and the
second column tells us that about 20 people watched between 1 and 2 hours per day (up to but
not including 2 hours) etc. Looking at this histogram we can see that most people watched
between about 2 to 6 hours per day. Histograms don't give us very specific information about
what's'typical'. We'll g et much more detailed infom画on when we look at some summary
statistics.
Histograms give a very good overall impression of the shape of the distribution. You can see
that there's one central peak and the hours of TV are fairly evenly spread on each side of this
peak. We say the distribution is approximately symmetric. Distributions can have all sorts of
different shapes:
10
of anxiety, but there was another large group of
。 students with high levels of anxiety.
0 10 20 30 40 50 60
1 Note that the histogram in Figure 1.5 has been modified to have an interval width of 1 and an 'anchor'ofO.
14
Copyright © 2016 Pearson Australia (a division of Pearson Australia Group Pty Ltd) ISBN (978148861 3432), Dr. Andi Garing, Foundations of Statistics, 4th Edition
Module 1: Critical Thinking in Statistics
2500
The histogram in Figure 1.8 represents house prices,
K3C
anbaJ:l75
in thousands of dollars. The tail end drags off to the
right, with a few unusually high values (outliers).
50250
We call this'positively skewed'.
Let's look at just one more. Using the data in the What_Analysis.sav data file, produce a
histogram for the variable'Time to travel to work (minutes)'. SPSS produces the histogram in
Figure 1.9.
25
20
AouanbaJ:l
1510 Note that the distribution is approximately
symmetric
50
10 15 20 25 30
Time to travel to 叩rk (Minutes)
Figure 1.9: Distribution of travel times.
40
If you are not using Version 23 of SPSS you 302010
,<;::,uanbaJ.::t
may get a somewhat different looking
histogram for this distribution, like the one in
Figure 1.10. This histogram looks like it has
three peaks. However, this is just a result of the
inappropriate way SPSS has grouped the values
。
15
Copyright©2016 Pearson Australia (a division of Pearson Australia Group Pty Ltd) ISBN (9781488613432), Dr. AndiGaring, Foundations of Statistics, 4th Edition
Module 1: Critical Thinking in Statistics
40
8 20
505
11
否 U$beLL
3020 64
A:luan
A:iuanbw:1
』
LbLa
2
10
。
。
。
10 20 30 40 50 60 4.0 4.5 5.0 5.5 6.0 6.5 7 .0 30 40 50 60 70 80 90
Feeling of well being ENERGY/ENTHUSIASM marks
5 11
505
芒
uanbaJ:t
33』
8d
50
。
5 10 15 20 0 11 12 13 14 15 16 17 18 19 20
Number of basketball matches Number of basketball matches
attended last year attended last year
Histogram Bar Chart
16
Copyright©2016 Pearson Australia (a division of Pearson Australia Group Pty Ltd) ISBN (9781488613432), Dr. AndiGaring, Foundations of Statistics, 4th Edition
Module 1: Critical Thinking in Statistics
There is a much more reliable way of indicating what the typical travel time is. You've
probably met it before - the mean. To calculate the mean, you just add up all of the values and
divide by how many cases you had. So, for example, suppose I asked 5 people how long they
took to travel to work, and their answers were: 18 min, 20 min, 22 min, 20 min and 20 min.
Then the mean would be :
18 + 20 + 22 + 20 + 20 100
Mean= =— = 20
5 5
So in this small sample the average time taken to travel to work was 20 minutes. This is easy
when you have just 5 cases, but it can be pretty tedious if you are dealing with hundreds of
cases. Fortunately for us, we don't need to do any calculations by hand at all —SPSS will do
all of the calculating for us. Part of the output we get from the EXPLORE procedure (see
Appendix A3, page 226) is a table of summary statistics:
Descriptives
Std.
Statistic Error
Time to travel to Mean 19.98 .318
work (Minutes)
95% Confidence Lower Bound 19.35
Interval for Mean
Upper Bound 20.61
5% Trimmed Mean 19.94
Median 20.00
Variance 25.249
std. Deviation 5.025
Minimum 8
Maximum 33
Range 25
Interquartile Range 7
Skewness .102 I I .154
Kurtosis -.259 .307
For the moment we're only interested i a few of these statistics, and I'磷highlighted them in
the table. In this sample, the aver~ e taken to get to work was 19 .98 minutes. There is
another statistic here which weals ells us what's'typical' —the median. The median tells us
that half of the respondents spent 20 minutes or less travelling to work1. In this case the mean
and the median are very similar. This is because the distribution is symmetric.
So we've got a couple of statistics that tell us where the distribution is centred, but we might
also like some infom画 on about how spread out the travel times are. The simplest thing to
look at here would be the smallest and largest values. These are given as the minimum and
maximum in our table of descriptive statistics. The people in this sample took between 8
minutes and 33 minutes to travel to work each day.
芦 statistic for measuring spread, called the variance, and a related
statistic, called the standard eviation. The variance takes into account all of the data, not just
the two end points. It measures how much the values vary about the mean.
1 To calculate the median by hand you would start by writing the values in order from lowest to highest. The
median is then the middle value. So in the example with 5 cases given above, the values ordered from lowest to
highest are: 18, 20, 20, 20, 22. The median is 20; the middle value. If there are an even number of cases, then the
median is half way between the two middle values.
17
Copyright ©2016 Pearson Australia (a division of Pearson Australia Group Pty Ltd) ISBN (9781488613432), Dr. AndiGaring, Foundations of Statistics, 4th Edition
Module 1: Critical Thinking in Statistics
For the mathematically minded, here's how it's calculated for the small sample of 5 cases we
had before. You won't need to calculate the variance by hand (that's what SPSS is for!) but
seeing how it's calculated can help you to understand what it's measuring.
So the variance is calculated by looking at how much each individual score differs from the
mean; squaring those differences, and averaging them. Basically it's a measure of how much
the scores vary about the mean. The descriptive statistics table gives both the variance, and
the standard deviation. The bigger the standard deviation (or the variance) the more spread out
the distribution. In our travel time example, the standard deviation, given in the Descriptives
table, is 5.028.
Descriptives
Statistic Std. Error
marks Mean 63.34 1.144
95% Confidence Lower Bound 61 .07
Interval for Mean Upper Bound 65.61
5%Trimmed Mean 63.32
Median 63.00
Variance 126.956
Std. Deviation 11 .267
Minimum 36
Maximum 90
Range 54
18
Copyright ©2016 Pearson Australia (a divisionof Pearson Australia Group Pty Ltd) ISBN (9781488613432), Dr. AndiGaring, Foundations of Statistics, 4th Edition
Module 1: Critical Thinking in Statistics
How do we interpret these percentiles? Start with the 10th percentile, 13 minutes. This tells us
that only 10% of the people in our sample took 13 minutes or less to travel to work. The 25th
percentile tells us that 25% of the people in our sample took 17 minutes or less to travel to
work, and so on. So the median, which we introduced earlier, is the 50th percentile.
There's a bit more new terminology to learn here. The 25th percentile is also referred to as the
first'quartile ' and the 75 th percentile is also called the third quartile - one quarter of the
scores are lower than the first quartile and three quarters of the scores are lower than the third
quartile.
Percentiles are very useful for placing an individual case within a distribution. So, for
example, if your travel time was at the 90th percentile, it would indicate that relative to other
people you took a long time to get to work. In fact, only 10% of people took longer than you
did.
Using percentiles can also help us to summarise the distribution - which leads us to our last
graph - the boxplot. Boxplots give a "5 number summary" of the distribution. They display
the minimum, first quartile, median, third quartile and maximum. The boxplot for travel times
is given in Figure 1.11.
35
30
25
——
24 min. ,~ —
,,
20 min.
20 Il一·—
~
l
17 min.
15
~ —·
10
5
Time totravel to work
(Minutes)
Figure 1. 11: Boxplot for travel times.
The numbers down the left hand side of the boxplot are travel times. The boxplot tells us at a
glance, that 50% of people in the sample took 20 minutes or less to travel to work, and the
middle 50% of travel times were between about 17 and 24 minutes. This gives us a feel for
both what is typical and for how variable travel times are. Boxplots don't contain as much
detail about the shape of the distribution as histograms, but they're really useful when you
want to compare several distributions.
19
Copyright©2016 PearsonAustralia (a divisionof Pearson Australi a GroupPtyLtd) ISBN (9781488613432), Dr. AndiGaring, Foundations of Statistics, 4thEdition
Another random document with
no related content on Scribd:
several thousands every year, but in 1902 they sank to the
comparatively small total of 600. But whilst every endeavour is made
to govern the country on lines acceptable to and understood by the
people, there are, of course, some points on which the policy of a
civilized government is necessarily in opposition to very deep-rooted
customs and habits to such a degree as to completely upset the old
basis of social life. In a minor degree this is true of the partial
application of the game laws to natives, but of far greater importance
is our attitude towards the institution of slavery and the slave-trade.
The occupation of the Soudan has been a tremendous blow to
slavery; one of the principal recruiting-grounds for slaves has
practically been closed. A certain amount of slave-raiding goes on
along the Abyssinian frontier. Descents are periodically made by
parties one to two hundred strong, well armed, from the south-
western districts of Abyssinia. They raid the Barun negroes, and
carry off the women and children. The same kind of thing is apt to
happen on the Darfur frontier, and some of the remote tribes in the
same quarter sometimes raid each other with the object of getting
slaves. Some of these, but not many, find their way to Dongola, or
the Ghezireh; others are taken to Tripoli. Special steps have now
been taken by the Anti-Slavery Department of the Egyptian
Government, which now has its headquarters at Khartoum, to put
down this traffic. Two extra English inspectors have been posted,
one at Rosaires, on the Blue Nile, the other at El Obeid. They are to
form small mounted corps of the best Arabs and patrol the disturbed
districts.
Apart from the actual work done, it will be a great thing to enlist
the best of the Arab tribesmen in the Government service, and it is
hoped that they will in time form the nucleus of an effective native
police. These men have been in the past some of the principal
exponents of slave-catching themselves; they ought to be very adept
at their new business. With these exceptions the slave-trade within
the borders of the Soudan has practically disappeared. During the
first year or two captures were occasionally made of small caravans,
but very heavy penalties were imposed. There is still a constant
demand for slaves in Arabia, and once a slave is shipped over the
Red Sea a good profit is assured. But it is too dangerous now for
anyone to try to make a regular livelihood by it. There are still,
however, about twenty cases a year of trials for offences against the
slavery laws, mostly isolated cases of kidnapping a woman or a
child, and probably there is besides a fair proportion of undetected
cases. But the regular trade is pretty well stamped out.
The benefits of the abolition of slave-raiding and kidnapping are
immediate and obvious to everyone. Even the Arab can understand
them, but he finds it very difficult to appreciate our attitude—which,
needless to say, is uncompromising enough—towards slavery as a
domestic institution. It is the one serious complaint which he has
against the new government. His domestic habits and customs have
been completely based on slavery for centuries. Slavery is permitted
and recognised by the Koran. In most cases the slaves themselves
have been treated more like members of the family than as slaves,
and no doubt many of them have had far happier lives than they
would have had in their own villages. Nor is it difficult to point to evils
which have arisen from the emancipation of the slaves. It is a
melancholy fact that many of the towns in the Soudan are crowded
with freed slaves, too lazy to do anything but steal, while the women
have recourse to an even less reputable occupation. It is easier to
break down the social system of centuries than to build up a sounder
fabric in its place. But the thing had absolutely to be done if the
Soudan was to have a real regeneration. Even when the slaves have
been well treated, the demoralization caused by slavery has been
great. The Arabs have all the vices of a slave-owning people. It was
a good time to make an absolutely fresh start. All changes of such
magnitude are bound to produce dislocations. The evils of the
change will die out with the present generation. The good must be
waited for patiently, but it is sure to come.
CHAPTER XIX
EDUCATION AND THE GORDON COLLEGE
Progress has not been very rapid. Lack of money and lack of
competent schoolmasters sadly hamper all operations. But kuttabs
are now established, attached as a sort of junior class to the schools
at Khartoum, Omdurman, Halfa, and Suakin. A model kuttab has
been established at Berber, which is reported to be doing well, and
another is being built at Dongola. The like is also being attempted at
Wad Medani, a populous town on the Blue Nile, capital of the
province of Sennar, with about 40,000 inhabitants. Reference has
been made to the lack of trained schoolmasters. Egypt itself feels
this difficulty, and Egypt is at present the only source of supply on
which the Soudan can draw. It was to meet this demand—at least,
so far as the kuttabs are concerned—that a small training college for
native sheikhs was opened in the beginning of 1901 in connection
with the school at Omdurman. At first this interesting experiment was
not very successful. The students, who all belonged to the best Arab
families, were all proud, ignorant, and lazy; and as Arabs they were
inclined to despise the Egyptian schoolmasters, whose task it was to
teach them. But now there is a great improvement. They have
increased in number to about thirty, and only lack of room prevents a
further increase. I watched them doing their own lessons, and also
receiving practical instruction in teaching by taking a class of the
school under the guidance of a master. It was impossible to doubt
the value of the experiments. They were nearly all fine-looking,
intelligent young men, some of them really handsome, with the keen,
clear-cut features that mark the pure-bred Arab. Three of them had
come from distant Kassala, where at present there are no means of
education whatever. The course lasts three years. At the end of it
they are to be examined as to their fitness, and they will then be
drafted off either to teach in their kuttabs or else to some posts in
connection with the native Courts. Whether as schoolmasters or
Cadis, they will be most useful elements in the development of the
Soudan.
In the negro portion of the Soudan, inhabited by the pagan tribes,
the people are so backward in civilization that the question of
education does not at present arise, or, if it does arise, assumes a
totally different aspect. Here is the field for the missionary. Two
missions are already established—one, the American Medical
Mission, on the Sobat; and the other, the Austrian Roman Catholic
Mission at Taufikieh, on the White Nile. Both are doing good work,
and both are to be encouraged and assisted by the Government. In
other parts of the Soudan it must be remembered that we are
dealing with a fanatically Mohammedan population, and any
suspicion that the Government was trying to proselytize would
immediately wreck all schemes of education, and probably be the
signal for grave disorders.
It will be a long time before the schools turn out sufficient pupils to
fill the Government Civil Service, and there does not seem to be any
danger of producing mere ‘babus,’ hanging about and relying on a
certain knowledge of English to procure them a job. The teaching of
English is entirely confined to those boys who are going to make use
of it in the Government Service or in commercial pursuits, where its
knowledge is required. For the ordinary mass of the population
nothing is to be gained by an imperfect knowledge of English. The
authorities are unquestionably right in discouraging such teaching;
the supposed political advantages of it are small, if not entirely
imaginary. The Soudan can never be a real white man’s country; its
rulers must always be speakers of Arabic, and its people will do far
better to employ their time in more useful ways than struggling with a
foreign language.
The Gordon College is the centre of education, but even now its
activities are not confined to mere teaching. It is proposed to form a
collection of books dealing with the Soudan, its peoples, its natural
history, and its various productions, accompanied by specimens to
illustrate them, and some progress has been made. Mr. Wellcome’s
valuable bacteriological research laboratory is in full working order,
and a skilled expert from Scotland has been in charge of it for some
months. He is doing a work of great importance, not only to the
Soudan, but also to the scientific world in general. The Soudan is a
land where strange diseases both of men and animals abound.
There is a wide field for research. The scientific and systematic
examination of these obscure subjects is already bearing fruit, and
cannot fail to ameliorate the conditions of life in these tropical
regions for the European as well as for the native.
Lord Cromer has promised that more shall be done for education
in the future. There is no need for hurry; indeed, it is essential that
the educational system shall be built up slowly with caution and
patience. But in time each province will have its own primary and
technical schools, whose pupils will be selected from the elementary
kuttabs. Afterwards, when all this has been carefully organized, the
provincial schools will in their turn pass on their more promising
students to Khartoum to receive the higher education which will then
be demanded. Then the Gordon College will at last become a real
college. In it the germ of a most hopeful future is contained. If the
work proceeds on the same sound lines as hitherto—which there is
no reason to doubt—great days are coming. The Arab is capable of
a very high degree of civilization, and has a great intelligence, which
has as yet had no chance of development. And in those days, I
doubt not, those who founded the Gordon College, and first kindled
the fire of learning, will be praised by its students as sincerely and as
deservedly as the benefactors of any of our own Western
foundations.
CHAPTER XX
‘I recently rode from Wad Medani on the Blue Nile to opposite Duem on the
White Nile, eighty miles across a perfectly flat plain sown almost throughout its
entire length with dhurra, which was standing 6 to 8 feet high. As there is only one
crop sown during the short rainy season, and as this is planted and harvested
within a period of sixty to eighty days, it follows that, if a system of irrigation were
possible in the Ghezireh it would become a huge granary capable of supplying, not
only the whole Soudan, but other countries as well.’
But, of course, any schemes for utilizing the waters of the Nile
have always to be considered in reference to the prior claims of
Egypt. It is only to Egypt that the Soudan can look for the money
necessary to carry out great works, and, naturally, Egypt would not
allow irrigation to be developed in the Soudan unless her own needs
were amply safeguarded. Whenever Egypt undertakes the great
works contemplated on the Upper Nile, the Soudan will share in the
benefit. That will not be for a long time to come, and meanwhile the
Soudan has opportunity to develop her resources and her population
so as to be able to seize the chance when it comes. There seems to
be no reason, however, why works on a small scale should not be
undertaken, subject always to the question of expense, to utilize in
the Soudan, by means of basin irrigation on the Blue Nile or the
Atbara, some of the flood-water during July, August and September,
which is only an embarrassment to Egypt now that perennial
irrigation is adopted so universally in that country. It is only in the
preceding summer months, when the Nile is low, that Egypt is forced
to watch the proceedings of her neighbour with such a jealous eye. It
would be rash to prophesy exactly when irrigation works will be
undertaken, or what form they will take; but it is certain that they will
come, and when they do, their effect upon the Soudan will be
immense. The subject is of peculiar interest to England; the lands in
question are capable of producing other crops than cereals, and, as
will presently be shown, for none are they more suitable than for
cotton.
Quite apart from any such speculations upon the future, the
external trade of the Soudan is capable of great expansion under
present conditions. An examination of the returns of imports and
exports during the last years before the rebellion is a very useful
guide to the capacities of the country. The figures are taken from
Colonel Stewart’s report.
There are no returns of imports except for the port of Suakin, but
these embrace practically the whole. Of course, a certain amount
came in by the Nile Valley route, just as now a certain amount comes
in by Suakin instead of by the railway. The position has been
practically reversed, and for purposes of comparison it is fair to take
the imports formerly entering the port of Suakin and those now
carried by the railway. The most important item was cotton goods,
which amounted to about 25,000 to 30,000 kantars annually. Linens
were about 200 to 300 kantars, and woollens 100 to 300 kantars.
There was also a certain amount of silks and silk thread and sewing
cotton. The Indian trade brought in a good deal of grain and tobacco.
In 1881 the import of Indian rice amounted to 20,000 kantars.
Petroleum (6,000 kantars), oil, zinc, copper, and iron, appear in the
list, as well as flour and provisions of various kinds in large
quantities, and candles, boots, and clothes. The peculiar tastes of
the Soudanese in luxuries are reflected in the large imports of
sandal-wood, and scents and perfumery, especially fish-scent (700
tons, or 15,400 kantars, a year). Both men and women are
particularly fond of strong, greasy scents. In all native festivities and
entertainments these scents are a very prominent feature, and a
native marriage can be smelt a very long way off. Last, but not least,
comes a very peculiar item: in one year no less than 11,000 dozen
umbrellas were imported. There is certainly enough sunshine in the
Soudan, and in parts of it rain also, to justify a large number of
umbrellas, but it is difficult to believe that 132,000 umbrellas would
find a ready sale in the Soudan to-day. Perhaps the Khalifa
particularly disliked the umbrella-carrying class.