[bit.do/azrs3] This is the 3^{rd} part of a 4-part lecture about an online course “Real Statistics: An Islamic Approach”. Anyone can register at any time, and complete the sequence of lessons offered at their own pace. The course is broken into independent modules, and the first module on general principles of Islamic Education is recommended for all Muslim teachers/students; to register, fill in form: Registration PIE.

Previous two parts of this lecture were Part I – Fundamental of an Islamic Approach {bit.do/azrs1} and Part II: Teachings of Statistics as an Act of Worship {bit.do/azrs2}. This part provides a new and different foundation for the entire subject. We will treat Statistics as a type of modern rhetoric, a way of persuading other people to believe in some idea using numbers as a tool for persuasion. This is radically different from the standard approach which treats statistics as an objective analysis of numbers. Link to the original 23m Urdu lecture at Univ of Baluchistan: RS (3/4) Stats As Rhetoric (urdu). The 21m English Video lecture is followed by a written summary:

**Statistics as Rhetoric**: (1900 Word Summary)

The Western approach splits the theory and practice. The Field Expert has knowledge of the real world, and takes observations and measurements, and turns them over to the statistician. The statistician analyzes the numbers and gets results, which he gives to the Field Expert. The Field expert does not have to know exactly what the statistician did; he can use the results for his application. The Statistician does not need knowledge of the real world – he only analyzes the numbers. Statistics provides us with OBJECTIVE truths, and hence all people should agree to the results of a valid statistical analysis. It is this myth of objectivity of the numbers and the analysis which allows the separation between theory and practice — the objective facts are the same for all, and hence do NOT require field expertise. {See “Connecting Statistics to Reality” for the **Central Fraud** at heart of modern statisics: the pretence of objectivity}

The Islamic Approach rejects the idea that theory and application can be separated. The same set of numbers, arising from different real world context, will require different types of analysis. So the statistician must always analyze numbers in the context of the real world phenomena which generated the numbers. See My Journey from Theory to Reality for more details about this argument. The Islamic approach rejects the idea that numbers are objective measures of reality. As we will see, most numbers being analyzed involve subjective judgments. We take the point of view that Statistics is a branch of rhetoric. We need to learn how to make ARGUMENTS with numbers.

The key rhetorical strategy of conventional statistics is hiding of the subjective elements of a statistical analysis. Both the data being analyzed, and methods of analysis, involve HUGE numbers of Subjective Assumptions. Conventional statistical analysis pretends that numbers, and analysis, is objective and factual, no room is left for arguments and persuasion. In this course, we will bring out the hidden value judgments, so that different perspectives can be explored, in light of different values, while having the same set of numerical measurements.

The key insight here is that most numbers are MADE UP, and involve HUGE numbers of subjective judgments. There are TWO types of Numbers – Facts and Fictions. The factual and objective numbers are about the External Reality. For example, Number of trees in forest, Number of people in Pakistan, Rupee Income of People in Pakistan, Prices of different goods in different places, the Quantity of Carpets produced for export. These can all be counted by numbers, and the numbers actually count something which is present in external reality, and therefore is objective.

However most numbers which enter statistical analysis, especially in the context of economics are number fictions, not number facts. These numbers are computed using subjective decisions which represent values, but these are hidden in the analysis. Instead the numbers are presented as if they are just like number facts, and hence objective measures of external reality, to which all observers would agree. Here are some examples of numbers which are fictional: IQ of a person, Wealth of Pakistan, Value of the Rupee in terms of purchasing power, Inflation Rate, Quality of Universities, Quality of Research produced by a faculty. Since this point is never made in conventional statistical texts, which treat all numbers alike, we will explain further why these numbers are fictional, not factual.

Why is IQ a Fictional Number? Because what it is supposed to measure is a multidimensional and qualitative factor, which cannot be reduced to a single number. It is computed by using a score on a quiz, but the same person would make different scores on different types of quizzes. If ten different people were asked to evaluate the intelligence of some one person, they would come up with different numbers. There is no cross-check available to tell which of the ten different scores is correct, and which one is wrong. Next time student takes same test, he will get different score. Different tests will lead to even more different scores. There is not much correlation between grades and SAT scores, even less between success and IQ. NO REAL underlying STABLE factor is being measured

Real phenomena are often qualitative, unobservable, unmeasurable, and unquantifiable. For example, how much love I have in my heart for Allah T’aala is an important reality of this type. Even when there are measurable phenomena, they are multidimensional. For example, Intelligence is a complex multidimensional phenomenon which cannot be reduced to a single number. The simplest case where fictional numbers are used is when multiple valid measures are reduced to one. Since this is one of the key factors which differentiates our approach from conventional statistics, we look at this case a bit further.

Basic Principle: One number CANNOT provide information given by Two numbers. Example: Tailor measures – height, neck, chest, girth – to make shirt. BUT readymade shirts use one number. HOW? One number is based on AVERAGE size of population. The use of subgroups – S, M, L, XL – clarifies the point that one number CANNOT and DOES NOT capture the full information available to the tailor. It should be clear that these categories provide only rough approximations to body shapes of most people. From the one number for body-size, we cannot recover the information that the tailor uses to provide a much better fit.

Statistics is used to make fictions appear as fact. Creation of a single number to measure something which does not exist, makes this thing come into existence! This is a POWERFUL rhetorical strategy, because it creates a “fact” about something which only exists due to our calculations, and does not exist in reality. An important example of this is Inflation – this is a fictional number, no one number can describe the complex changes in prices which take place in the entire economy. Before we explain this in greater detail, it is important to *pause to explain how our subjective consensus creates realities*.

**Subjective Consensus Creates Realities**: A large part of our mental landscape consists of ideas which come into existence because we all agree to them. For example, Pakistan exists because it is recognized as a country by all of us. If tomorrow, all mankind comes to the realization that we are all brothers and sisters, sons and daughters of Adam AS and Hawwa AR, we could decide to abolish all nations and boundaries. Then all the countries would cease to exist. Not that this is a realistic scenario, but it is meant to clarify that countries exist because of our agreement, and they can cease to exist if we agree to abolish them. Note that even though countries do not exist in the external reality, they are of extreme real importance in our lives. So even “facts” which are created by subjective calculations can play a very important role in our lives. A very important example of this is the rate of “inflation”.

As already stated, Inflation is a fictional number. Why is inflation fictional? Because multiple numbers are reduced to ONE. Over the time period, in different regions, for different people, for different types of consumption patterns, the increase in prices could be very different. But it is all reduced to one number, because this makes it easier to understand. Once the number is announced, it becomes real. People make plans, and investment decisions on this basis. Monetary policy is conducted on this basis. If we reduced the level of fictionality, by separating inflation into different categories, and then looking at different classes of people, we would find many differences, and could make better plans. Instead of looking at overall inflation, identify sectors with surplus capacity and providing these sectors with money, while keeping money supply restricted in sectors where capacity limits have been reached and inflation is high, would provide us with better monetary policy. So use of fictional numbers has real impacts on our lives.

**Very Important Fact**: Western epistemology rejects the unobservables. While the Quran starts by describing those who believe in the unseen, influential Western philosopher Kant argued that we have no access to the unobservable realities EXCEPT by what we can observe about them. So, he argued (see Kant’s Blunder) that we should focus on what we can observe, and not worry about the hidden realities which are forever inaccessible. It is this mistake which leads to the current methodology of statistics and econometrics, which focuses on the numbers, and not on the hidden realities which lie behind the numbers. For the sake of clarity, we will refer to conventional statistics as “Nominal” statistics, because it is concerned with the “names”, that is the numbers used to measure observables. In contrast, “Real” statistics is always concerned with the complex and qualitative realities (which may or may not be observable) which generate the numbers that we see.

Now we can clarify the concept of fictional numbers: *Fictional Numbers measure things which do not exist in external reality (even though these things may exist by our collective agreement)*. We illustrate this abstract idea by some concrete examples:

What is body type (a single number which is enough for the tailor)? NONE – it does not exist.

What is the rate of Inflation? NONE – No one number can characterize the inflation rates.

What is the QUALITY of a University? NONE – No one number can capture quality.

What is the QUALITY of research done by a faculty member? There is no ONE number for it.

When ARTIFICIAL numbers are CREATED to measure unmeasurables – These HAVE impact on the REAL world

Quantify Research by number of publications in HEC recognized journals => huge increase in meaningless publications. Increase in fake journals.

Measure police performance by number of challans issued LED TO huge increase in fake challans for trivial offences.

Measure Inflation by a Laspayre Index: This involves averaging price increasing using (arbitrary) weights of actual consumption from one fixed base year. This can leads to Monetary Policy Decisions which may hurt MOST of the population. There are many other possible choices for weights, but all choices are arbitrary and subjective. Because the Wealthy have HUGE proportion of the consumption bundle, and hence weighting by the actual consumptions provides a more accurate picture of inflation as seen by the wealthy. In contrast, the Sensitive Price Index, measures more accurately the inflation seen by the poor. Generally these measures move together, so, often, not much is lost by considering one number. But the important point here is the conceptual one, that there is no one number which CAN measure inflation. More generally, complex multidimensional phenomena cannot be reduced to one number. The Laspayre index corresponds to an arbitrary choice of weights (using base year), while the Paasche index uses current year weights. To see how much difference this can make, look at “Which base year?”.

**Summary & Conclusions**: Most numbers used in statistical analysis are FAKE – they do not correspond to external realities. MANY subjective judgments go into creating these numbers – BUT they are presented as OBJECTIVE facts about external realities. Learning to differentiate between number facts and number fictions is very important in learning how to do real statistics. In addition, we also need to learn the boundaries of statistics. We need to learn how to deal with qualitative phenomena, which are not measurable, not observable. In Q&A session, the question was asked: How can we deal with unobservable qualitative phenomena? The answer is: we have a lifetime experience in doing so. We manage our daily lives, without measuring most of the things we do. We interact with many people, and judge their reactions to us, even though what is in their hearts is not observable.

Pingback: Number Facts & Number Fictions | WEA Pedagogy Blog

Pingback: Number Facts & Number Fictions | Real-World Economics Review Blog

Pingback: Real Statistics (4/4) Illusion of Objectivity | An Islamic WorldView

Pingback: Real Statistics: An Islamic Approach | An Islamic WorldView

Question from Bairuni Group:

Can we say that Conventional Statistics is just Black Box because it is just representing the very little proportion of reality and ignoring the real factors? That real factors are mostly qualitative in nature, so conventional or (Deceptive) Statistics have no capacity to represent the hidden truth.

This is pretty close to the mark – the diagram in the next post Real Stat (4/4) makes the point clearer, but you have expressed the same idea in words.

Dear Sir,

Many thanks for encouragement. I am pasting here link of book, which will helpful to understand the deceptive Knowledge.

Pingback: Q & A: Islamic Approach to Statistics | An Islamic WorldView

Pingback: Kelvin’s Blunder: Measure Everything | An Islamic WorldView

Pingback: The Struggle for Knowledge | An Islamic WorldView

Pingback: AZ Students

Pingback: Understanding Statistical Distributions | WEA Pedagogy Blog