Lecture -- 1-- Start - University of Pittsburghsuper7/51011-52001/51331.pdf · Quantitative Techniques for Social Science Research Ismail Serageldin Alexandria 2012 Lecture # 1: Science,

Lecture -- 1-- Start

Outline

1. Science, Method & Measurement2. On Building An Index3. Correlation & Causality4. Probability & Statistics5. Samples & Surveys6. Experimental & Quasi-experimental Designs7. Conceptual Models8. Quantitative Models9. Complexity & Chaos10. Recapitulation - Envoi

Outline

1. Science, Method & Measurement2. On Building An Index3. Correlation & Causality4. Probability & Statistics5. Samples & Surveys6. Experimental & Quasi-experimental Designs7. Conceptual Models8. Quantitative Models9. Complexity & Chaos10. Recapitulation - Envoi

Quantitative Techniques for Social Science Research

Ismail SerageldinAlexandria

2012

Lecture # 1:Science, Method, and

Measurement

On Science

On Science

Science Is Driven by Curiosity About the Natural World

WHAT IF?

WHAT IF?

WHAT IF?

??

Defining Science

“The organization of our knowledge in such a way that it commands more of the hidden potential in nature ..”

J. Bronowski

Intellectual Activities

NaturalSciences

PhysicsChemistryAstronomy

GeologyBiology

Etc.

HumanSciences

PsychologyEconomics

Political ScienceSociology

HistoryEtc. Applied Fields

TechnologyEducationMedicine

LawEtc.

Humanities

EstheticsEthics

ReligionPhilosophy

Etc.

Sciences

Curiosity

Influence others

Objective

Subjective

Natural Sciences (classical definitions)

• Physical Sciences : Physics, Chemistry

• Life Sciences : Biology (zoology, botany)

• Earth Sciences : Geology, Astronomy, Meteorology

Overlapping Domains in Science

Biochemistry, Paleontology, Molecular Genetics…

Changed Outlook: Process and System Views

Example: Photosynthesis

• Light: the Energy source (physics )

• Photosynthesis : The food productionprocess (chemistry )

• For plants (biology )

Energy…biochemical pathways…cell Biology… plant physiology…

The Nature of Scientific Knowledge

• Falsifiable (Popper)• Approximative• Empirical• Replicable

• And so much more…

Karl Popper(1902-1994)

Before scientific thinking can proceed, certain philosophical presuppositions must be made about the nature of the universe:

• Objective reality exists – there really arethings out there, everything is not simply afigment of the imagination .

• The universe is knowable – no aspects of theuniverse are beyond human understanding .

• The universe’s operation is regular andpredictable – if events occur at random,without any warning or pattern, no amount ofanalysis will uncover any regularity to them .

Philosophical Presuppositions













On The Scientific Method

The Scientific Method

The Method of ScienceOBSERVATION: Sense specific physical realities or events.

HYPOTHESIS: Create a statement about the general REVISEDnature of the phenomenon observed. HYPOTHESIS

PREDICTION: Forecast a future occurrence PREDICTIONconsistent with the hypotheses.

EXPERIMENT: Carry out a test to see if predicted EXPERIMENTevent occurs.

If results DO If results DO NOTmatch prediction, match prediction RECYCLEhypothesis issupported.




















If results DO If results DO NOTmatch prediction , match prediction RECYCLEhypothesis issupported.





If results DO If results DO NOTmatch prediction , match prediction RECYCLEhypothesis issupported .


























No amount of experimentation can ever prove me right; A single experiment can prove me wrong.

Albert Einstein

The Scientific Method

• Conjecture• Hypothesis• Testing• Replicability• Falsifiability• Cumulative evidence• Explanatory power• Predictive power

Math

• Logical, consistent, proof is absolute within its own axiomatic rules

• Math is added to, science is replaced• Math is the science of patterns• It is elegant, beautiful and concise• It demands enormous precision in

thinking clearly about abstract objects

Math and Science

• Enormous power for manipulating quantitative results

• Hence questions of measurement are important

• Quantification and qualitative analyses remain important issues

Back To The Social Sciences


NaturalSciences


GeologyBiology

Etc.

HumanSciences

PsychologyEconomics




LawEtc.

Humanities

EstheticsEthics

ReligionPhilosophy

Etc.

Sciences

Curiosity

Influence others

Objective

Subjective


NaturalSciences


GeologyBiology

Etc.

HumanSciences

PsychologyEconomics




LawEtc.

Humanities

EstheticsEthics

ReligionPhilosophy

Etc.

Sciences

Curiosity

Influence others

Objective

Subjective

Back To The Scientific Method

The Method of Science

Observation

Hypothesis

Prediction

Experiment

The Method of Science

Observation

Hypothesis

Prediction

Experiment

Interpretation

But can we apply that empirical method in the social sciences?

Understanding Society To Design

Social Policies

Policy Counts

Economic Growth and Poverty Reduction

Growth and Poverty Reduction

• Growth is a necessary but not sufficient condition for poverty reduction

• The quality of growth and the nature of the policies matters enormously

0.0

2.0

4.0

6.0

8.0

10.0

10.08.06.04.02.00.0 12.0

% annual growth in GDP/person

% annual decline in poverty(Headcount

index)


0.0

2.0

4.0

6.0

8.0

10.0

10.08.06.04.02.00.0 12.0



index)


0.0

2.0

4.0

6.0

8.0

10.0

10.08.06.04.02.00.0 12.0



index)

Jamaica

Costa Rica

India

MalaysiaIndonesia

SingaporeThailand

Taiwan

Sri LankaMexico

Brazil

Bangladesh

Pakistan


0.0

2.0

4.0

6.0

8.0

10.0

10.08.06.04.02.00.0 12.0



index)

Jamaica

Costa Rica

India

MalaysiaIndonesia

SingaporeThailand

Taiwan

Sri LankaMexico

Brazil

Bangladesh

Pakistan


0.0

2.0

4.0

6.0

8.0

10.0

10.08.06.04.02.00.0 12.0



index)

Jamaica

Costa Rica

India

MalaysiaIndonesia

SingaporeThailand

Taiwan

Sri LankaMexico

Brazil

Bangladesh

Pakistan


0.0

2.0

4.0

6.0

8.0

10.0

10.08.06.04.02.00.0 12.0



index)

Jamaica

Costa Rica

India

MalaysiaIndonesia

SingaporeThailand

Taiwan

Sri LankaMexico

Brazil

Bangladesh

Pakistan


Policies, Inequality and Welfare

Is Inequality Built Into Economic Structure?

• Is movement into knowledge based economy necessarily accompanied by inequality?

• Is US economy intrinsically generating more inequality?

Poverty Observed: US and Selected European Countries, 1991

0 5 10 15 20 25 30

Belgium

Canada

Denmark

France

Ireland

Italy

Netherlands

Sweden

UK

USA

% POORSource: Robert Solow, “Welfare: The Cheapest Countr y”’in NYRB, 23 March 2000, p. 20-23

Poverty before Government policy effects

US and Selected European Countries, 1991

0 5 10 15 20 25 30

Belgium

Canada

Denmark

France

Ireland

Italy

Netherlands

Sweden

UK

USA


Policy Effects on Poverty:Pre and Post tax and transfers, 1991

0 5 10 15 20 25 30

Belgium

Canada

Denmark

France

Ireland

Italy

Netherlands

Sweden

UK

USA



0 5 10 15 20 25 30

Belgium

Canada

Denmark

France

Ireland

Italy

Netherlands

Sweden

UK

USA



0 5 10 15 20 25 30

Belgium

Canada

Denmark

France

Ireland

Italy

Netherlands

Sweden

UK

USA



0 5 10 15 20 25 30

Belgium

Canada

Denmark

France

Ireland

Italy

Netherlands

Sweden

UK

USA


Why Quantitative Analysis ?

The Importance of Social Research

Much Economic Analysis Erases the Human Factor

The Need for Social Inputs Into Development Decisions

• Social policy is more than the social consequences of economic policies

• Social goals and policies complement economic ones

• Economic Analysis by itself is insufficient: Social, cultural, political and ethical dimensions must be introduced

Elements Of A Social Policy - I

• To maintain social cohesion• To foster equity

• To reach the ultra poor and other marginalized groups

• To uphold cultural identity (shared universal values and solidarity, not divisive micro -identities)

Elements Of A Social Policy - II

• To promote participation (voice, choice and empowerment through access to knowledge and resources)

• To facilitate social mobility (inter-generational, geographic and occupational)

• To support institutional development• To enable participatory social

research

Participatory Social Research

• Promotes more effective understanding

• Leads to sounder policy and program designs

• Empowers the people participating in the research

Social Research and Decision Making

• Coherent Framework

• Predictive

• Prescriptive



• Predictive

• Prescriptive



• Predictive

• Prescriptive

Crisis in the Non -economic Social Sciences

• Absence of theoretical framework for the dynamics of social change

• The negative impact of the post-modern currents

• Confusion about quantitative and qualitative aspects pf research

• The misunderstood role of models




• Confusion about quantitative and qualitative aspects pf research





• Confusion about quantitative and qualitative aspects of research








• Absence of a theoretical framework for the dynamics of social change




Quantitative Social Analyses:Laplace

“Let us apply to the political and moral sciences, the method founded on observation and mathematics that has served so well in the natural sciences.”

-- Pierre Simon de Laplace

(1749-1827)

Quantitative Social Analyses:Quetelet

“The more advanced the sciences have become, the more they have tended to enter the domain of mathematics, which is a sort of center toward which they converge. We can judge of the perfection toward which a science has come by the facility, more or less great, with which it may be approached by calculation.”

-- Quetelet (1796-1874)

Quantitative Social Analyses:Quetelet

Quetelet (1796 -1874), by the way, invented the notion of the “average man.”

Quantitative Social Analyses:Boorstin

“Today, the Cassandras of social science speak the language of numbers”.

-- D.J. Boorstin (1914-2004)

Source: Daniel J. Boorstin, Cleopatra , (op.cit., p142)

Vehement Reactions

• Dehumanizing the humanities• Denies individualism• Treats people like products or

machines• Economics is not the whole story• Etc. etc.

Serageldin On Reductionist Views

Serageldin on Reductionist Views

• Three buckets of water and a handful of minerals held together by chemical reactions…

• A society is more than the sum of its economic and financial transactions…

+

≠

Serageldin on Reductionist Views

• Three buckets of water and a handful of minerals held together by chemical reactions…

• A society is more than the sum of its economic and financial transactions…

≠

Conclusions

• We need more, not less, sophisticated approaches…

• Clever word games are not helpful to either explain social realities or to help formulate polices and interventions that improve the well-being of people

But we need quantitative analysis to understand, and to

measure and to devise appropriate Social policies

Quantitative Analysis complements Qualitative

analysis and frequently under-girds it.

On Measurement

Measurement counts

Measurement Is Important

• We treasure what we measure• Prescription and dosage depend

upon accurate estimation of magnitudes

• Establishing trends is as – or more --important than snapshots of magnitudes

• Monitoring of progress over time





















143/xxx

On Measurement

• Accuracy & Precision

• Resolution & Randomness

• Types of Scales

144/xxx

On Measurement



• Types of Scales

Accuracy in Measurement

• Using the right tool• The quality of the tool is important• How carefully we measure with it is

also important• Let’s use a ruler to measure the

length of a piece of wood…


• Using the right tool• The quality of the tool is important



also important



also important• Let’s use a ruler to measure the

length of a piece of wood…

Accuracy & Precision

• Accuracy : how close the measured value is to reality (i.e. what it ought to be) -- So if the ruler is defective and two rulers yield different results that is an error of accuracy

• Precision : is a measure of the reproducibility of the measurement,

Accuracy & Precision

• Precision : is a measure of the reproducibility of the measurement, our confidence that uncertainty of measurement has been reduced to a minimum.

• Sometimes the problem is instrumental precision (level of resolution) or the randomness of the event being measured.

On Measurement



• Types of Scales

Resolution vs. Randomness

• We do not re -measure the piece of wood 100 times and take the average.

• Assuming the wood was measured carefully, the error here is due to the resolution of the ruler, not the randomness of the event being measured.

Resolution of the tool

• So, instead of a ruler use higher resolution instrument s like precision Verniercalipers:

But that is very different from dealing with random events

Random events

Random events/outcomes require a probabilistic treatment

Social Science studies of events/outcomes usually require a statistical probabilistic treatment

Here multiple measurements and probabilistic techniques are used

We will get back to Probabilities Later

On Measurement



• Types of Scales

What kind of scale do we use in Measurement?

Four Kinds (Types) of Scales

• Nominal

• Ordinal

• Interval

• Ratio

Types of Scales:I. Nominal Scales

Numbers are used to name, identify or classify.

The onlypermissible arithmetical

procedures are counting and

statistical techniques based

on counting. Level Limitations Example

163

Social Science Examples of Nominal Scales

• Marital Status: Married, Unmarried• Nationality: Chinese, American,

European, Egyptian• Religion: Muslim, Christian, Jewish,

Buddhist, …• Ethnic or tribal group • Race: Black, white• P/F Evaluation: Pass/Fail

Types of Scales:II. Ordinal Scales

Numbers indicate rank or order.

Ranking methods and other statistical techniques based on

interpretations of “greater than” or “less than” are

permissible.

Level Limitations Example

165

Social Science Examples of Ordinal Scales

• Grading of interpersonal skills

• Evaluating managerial skills

• You can say greater than, but you cannot really quantify the amount or degree objectively.

Types of Scales:III. Interval Scales

The intervals or distances between each number and the next are equal, but it is not known how far any of them is from zero.

Addition and subtraction and statistical

techniques based on these two operations are

permissible. Multiplication and di-

vision are not permissible.


167

Natural Science Examples of Interval Scales

• Temperature:– Two days: 20 and 40 degrees Celsius– Difference between them is 20

• Cannot say twice as hot because zero could be:– Celsius scale– Fahrenheit scale– Kelvin scale

Social Science Examples of Interval Scales

• Grading school exams: • Say two students took a test: results

score was 20 and 40 points (difference is 20 points) but should not say twice as much.

• BUT…teacher could have added a few easy questions that would have obtained each student 10 more points

• Results would have been 30 and 50 .

Social Science Examples of Interval Scales

• Height or weight of people:– Say two persons 1.6 m and 1.80 m– Or two persons weigh 60kg and 80 kg

• Note:– No one is really 0.0 height or weight

Types of Scales:IV. Ratio Scales

Each number can be thought of as a distance measured from zero

There are no limitations. All arithmetical operations and all statistical techniques are permissible


171

Social Science Examples of Ratio Scales

• Income and expenditure • Years of schooling• Number of respondents selecting

something• Number of respondents who have a

particular cardinal quality (e.g. married, unmarried)

• Etc.

Social Science Examples of Ratio Scales

• Income and expenditure • Years of schooling• Number of respondents selecting

something• Number of respondents who have a

particular cardinal quality (e.g. married, unmarried)

• Etc.

It is important that not all relationships or all mathematical operations can be applied to all

scales.

Now let’s move on to some descriptors of groups

What Is An Average?

We say: I examined 20 students and the average score was x

What is the meaning of the word “average”?

Average?

• Mean: usually add up the values for all the observations and divide them by the number of observations

• Median: the number at which half the observations are smaller and the other half are bigger

• Mode: the number that appears most frequently in the distribution of observations.

Average?




Lets take 20 observations

• 1, 2, 3, 4, 4, 4, 4, 4, 4, 4, 6, 6, 6, 6, 7, 7, 8, 10, 10, 20

• What is the Mean ? • The Median ? • The Mode?

Lets Find the Mean

• 1, 2, 3, 4, 4, 4, 4, 4, 4, 4, 6, 6, 6, 6, 7, 7, 8, 10, 10, 20

• Mean =Total / number of observations

• Total = 1+2+3+…. + 10+10+20 = 120• Mean = 120 / 20 = 6

Formula for the Mean

n = ��⋯��

�

183

Lets Find the Median

• 1, 2, 3, 4, 4, 4, 4, 4, 4, 4, 6, 6, 6, 6, 7, 7, 8, 10, 10, 20

• Median = 5

Lets Find the Mode

• 1, 2, 3, 4, 4, 4, 4, 4, 4, 4, 6, 6, 6, 6, 7, 7, 8, 10, 10, 20

• Mode = 4

So for these Observations

• 1, 2, 3, 4, 4, 4, 4, 4, 4, 4, 6, 6, 6, 6, 7, 7, 8, 10, 10, 20

• Mean = 6• Median = 5• Mode = 4

Lets Change one Observation

• 1, 2, 3, 4, 4, 4, 4, 4, 4, 4, 6, 6, 6, 6, 7, 7, 8, 10, 10, 1000

• Mean = 1100 / 20 = 55

• Median = 5• Mode = 4

Introducing: The Normal Curve(The Bell Curve, The Gaussian Distribution)

Mean, Median, Mode

Average?




Mixing between these is one of the most common fallacies in

reporting social statistics

We will come back to the Normal Curve(The Bell Curve, The Gaussian Distribution)

many times in this course

But that will be for later…

SO…

Let’s make sure that we keep our heads above water!

I want you all to be swimming, not drowning!

Then we will all learn to fly…

Thank You

Lecture -- 1-- Start - University of Pittsburghsuper7/51011-52001/51331.pdf · Quantitative Techniques for Social Science Research Ismail Serageldin Alexandria 2012 Lecture # 1: Science,

Documents