On the contrary, having multiple sources of data can be healthy for the measurement and monitoring of MDGs at country level. Validation is a basic tool for ensuring data quality and yet there is no better way of data validation than either a repeated survey or use of a similar or multiple data sources. A repeated survey seems a more costly venture than having a compromise amongst multiple data sources measuring the same indicator. Basically, sample surveys are designed to meet specific objectives, but it is also true that certain sample surveys may have the same specific objectives; hence the likelihood to yield more reliable data when used for validation purposes to improve data quality for the timely monitoring of MDG indicators.

 Data discrepancy therefore poses a more fundamental challenge to researchers in the field of statistics to develop more reliable methodologies for detection of data discrepancies and align them accordingly. Data discrepancy management systems (DDMS) can then be proposed and developed to handle this eminent challenge to monitoring of MDGs in Africa as we tend towards the year 2015.

Kindly give your opinion below…

During the presentation at the workshop on the development and management of MDGs databases at the country level held in Nairobi, Kenya for 15^th to 19^th February, a presenter in his analysis gave model-based estimates as one of the causes of discrepancies between national and international data. Whereas it might hold some drops of water, there is no sufficient evidence to believe that the national estimates are better given the fact that they suffer from very high sampling errors of the magnitude higher than 11 percent with low study coverage, bias error and poor indicator specifications. Many developing countries are faced with scarcity of sufficient and consistent data to monitor MDGs. Besides, between the year of MDG declaration, 2000 and 2005, no data, hence no report was published showing the status of MDGs in any African country. For example, in Uganda, the first ever substantive MDG report was released in the year 2007 and as I write, the second MDG country status report is yet to be released. Therefore, under such circumstances, shouldn’t model based data be used to facilitate monitoring of MDGs at country level? What about situations whereby resources both financial and human resource are insufficient for carrying out reliable surveys and yet the country or the international community need to know its MDG status? On the positive note, what if the country has good statisticians that can develop data based models whose sensitivity analysis show very high correlations between expected data and model-based data?

In conclusion, model based data are increasingly becoming an integral part in the monitoring of MDGs in the both developed and developing countries, but more so in the developing countries majorly because of their unsustainable ability to regularly collect quality data over time.

What is your opinion?? Kindly give your opinion below…

In the previous posting, we made an attempt to understand what statistics is all about. Thanks for the compliments from all readers especially about the article ‘statistics aren’t just figures’, may God bless you to read even more!

Let’s make an attempt to define statistics without reservations. Statistics is all encompassing and therefore, it would be egotistical to restrict its definition to only quantitative aspects. In the beginning when the author of this article was growing up as a young statistician in a remote village, his teachers in primary school by then sarcastically didn’t know the existence of statistics. They simply knew mathematics, and other subjects like science and social studies! Someone has just told me here that even up to now, mention of statistics begins somewhere after primary education level. Corollary, the beginning of statistics in Uganda comes after a child is a teenager. Possibly, the curriculum developers think this is the right age for the child to appreciate the otherwise all inclusive subject of statistics.

Harris and Taylor (2009) in a forward of their book Medical Statistics Made Easy wrote “A love of statistics is, oddly, not what attracts most young people to a career in medicine and I suspect that many clinicians, like me, have at best a sketchy and incomplete understanding of this difficult subject.”

According to the American Statistical Association, Statistics is the scientific application of mathematical principles to the collection, analysis, and presentation of numerical data. Statisticians apply their knowledge of statistical methods to a variety of subject areas, such as biology, economics, engineering, medicine, public health, psychology, marketing, education, and sports.

A more open definition of statistics is given by the department of statistics of the University of Nebraska, Lincoln whose motto is turning data into knowledge to solve real world problems. They define statistics as the development and application of methods to collect, analyze and interpret data. Modern statistical methods involve the design and analysis of experiments and surveys, the quantification of biological, social and scientific phenomenon and the application of statistical principles to understand more about the world around us.

An interesting definition is that statistics is the study of data. This implies that a study involving either quantitative or qualitative data is basically statistics.

Many issues arise from these definitions, but I will just conclude without delving into the discussion by pointing out few generic questions. Does statistics imply strictly quantitative data? Hence, is statistics restricted to the study involving strictly numeric data? Is quantification of qualitative data a function of a statistics? Thus, is qualitative study independent of statistics? 

The magnitude of statistics should not be limited by only what is taught in class, rather there is a great need to attempt looking outside the statistical black box. Participating in the workshop at Windhoek, Namibia from 3rd – 16th October 2007 that was organized and funded by the Inwent Centre for Economic, Environmental and Social Statistics was an eye opener of how wide and broad statistics can be. The theme was ‘On Measuring Democracy, Governance and Human Rights’ and managed by Dr. Uwe Singer. At first, I wanted to turn down the offer to attend reasoning that I have nothing to do with political issues since at no time have I ever dreamt of being even a local council one chairman and worse still, although it is my right, I have never voted! Agnes, Abraham and I thought we were going to learn more about Democracy, Governance and Human rights aspects, but instead we went through the application of statistics.

It was not long after introductions that Dr Ken Mease of the University of Florida and University of the West Indies, Trinidad introduced the Statistics Camp and this took the lion’s share of the time as he led us through surveys, data quality and measurements. Some of the survey types he introduced that of course seemed unique from what you learn from the sampling class included.

• Country reports
• Cross-comparative surveys
• Expert assessments
• Expert panel
• Household survey
• Mass opinion survey
• Media Indicators
• Business surveys

I guess you may be wondering the theoretical linkage of this kind of survey type classification, but the truth is that these types are generated from the practice of statistics. It does not matter whether it is you, the Statistician or otherwise practicing it, but comes as a result of the need to see statistics at work.  

The workshop advanced and stochastically need arose of practicing what we had learned. As participants put it, ‘talking statistics is one thing, practicing it and bringing it to live, is another – at times quite exhausting, but equally exciting issue.’. After ten days of theoretical and group work, the trainees were asked to split up into groups, put on their walking shoes and do a comparative household survey in three different parts of the Namibian capital, Windhoek: Klein Windhoek, a residential area; Windhoek West, with a lot of middle-class citizens and Katutura, an area consisting of poorer neighborhoods. ”Are you happy with the delivery of public services like water, sanitation, electricity and garbage collection?” was the overall question to ask the citizens of the different parts of the city.  “How do you consider the performance, fairness, responsiveness, participation and accountability in the provision of these services?”

What seemed to be an easy task at first glance, turned out to be challenging – for various and sometimes unforeseen reasons. For example, I was in the group that walked out to Windhoek West. We were very early and we were confronted with a late waking up in that area. We had to deal with vicious barking dogs of all shapes and sizes. We spent most of the first hour ringing several gate bells in vain and hoping that the fences were strong enough to restrain the barking dogs. Some of the residents, mainly of European descent, did not even realize that this was a household survey, they thought, Africans knocking on doors during a weekend are simply seeking for servants jobs whereas others were reluctant to answer the questions.

We learnt that every household survey should be announced and advocated properly in advance, in order to be most successful and effective. Despite these difficulties, the comparative data collection exercise was an enriching experience, according to the participants. They also appreciated that they were greatly involved in the data entry, analysis and interpretation. Donald, a lawyer and one of the Ugandan participants put it as “This was truly statistics in motion!” The interesting part however, was that I took over as a facilitator towards the end of the workshop as I handled the nasty parts of data management and analysis of the aspects of Measuring Democracy, Governance and Human Rights.

What is your view?

Defining statistics these days has become a statistic itself. The institute of statistics and applied economics, ISAE, the only and only recognized statistical training institute for the Anglophone speaking countries in the region and Africa as a whole is associated with the saying ‘using figures to make informed decisions’. So would you like to believe like others who read ISAE fliers and its website  ( http://www.statistics.mak.ac.ug ) that figures are synonymous to statistics? Maybe that could be one definition, but remember there should be only one correct definition else it is incomplete thus incorrect!

Alright at the recently concluded 57^th ISI conference, the first ever ISI conference (refer to http://www.wesonga.com ), that took place in Durban from 16^th to 23^rd August 2009 there was a display of papers on statistics. One motivating finding was that there were basically two types of papers presented, one was statistics and the other was application of statistics. Oh no, do not ask me yet what statistics is, but the conference reality was that in some sessions, a paper would be presented introducing a new theorem, its lemma and proofs! In the end one may or may not conclude with just an example, not case study findings – no figures at all!  So is theorem tantamount to statistics?

If we took an example of a statistic, the most popular one which even politicians know, known as the mean, a measure of central tendency and we translated it into something like the one below.

1. for (i=0; i <= N; i++)
2. {
3.             sum += X[i];
4. }
5. mean=Sum/N;

We should note that the given five statements above are referred to as code, but not figures. So is code equivalent to statistics?

Hopefully, we now know what statistics is else if not yet known then watch this space, but  you may also contribute for a more detailed discussion and definition of statistics.

Over to you!!

Plans were hatched and thanks to the Almighty, a declaration was made that Africa will host another first ever International Statistics Institute conference on the African continent at one of the most beautiful city of Durban still in the Republic of South Africa. This 57th conference is slated to commence on the 16th till 22nd August, I mean this month, remember today is the 1st August, 2009. Never has this ever happened in Africa, note none of all the 56 conferences held before! Actually the president of ISI (www.isi.cbs.nl) in her closing remarks of the International Young Statisticians conference wondered the level at which Young African Statisticians really displayed statistics during the presentations of their papers, remember stochastic model of statistical displays!! Actually, none of them thought there was going to be any significant methodological and applicable statistics at the conference, but alas, they got convinced and decided that we now host the ISI-2009 at Durban city port! For conference details, go to www.statssa.gov.za/isi2009.  So let’s fold our sleeves, unbutton our shirts and fold our trousers to get ready not only to attend, but to participate fully at the first ever International Conference in South Africa. Whether, it means walking, we will because yes, yes we can, go anywhere on our continent!

Other first ever cases include: first ever black US president, first ever woman ISI president, first ever woman ISAE Director, first ever Math Makerere University VC,…. to mention, but a few!!

But alas! The officials at the Japanese Embassy in Uganda, visa section denied me a visa. How i wish i was told the reason as to why i was denied a visa to travel to Yokohama, Japan to present a paper despite presentation of all the documentation required including, inevitably an invitation letter, hotel booking, travel arrangements throughout my course of stay and these were sent through mail to me to be delivered to the Japanese Embassy in Uganda! I didn’t have any intention to disappear or stay in Japan beyond the prescribed visa period! Let alone, I had my return air ticket, thanks to the generosity of the Civil Aviation Authority, Uganda for the sponsorship.

The paper programmed for presentation at the International Association of Statistical Computing conference is titled Statistical Models for Air Traffic Management: case study Entebbe International Airport. See programme for paper number 54.

Personally, I am busy searching for email addresses of the ISI president and Dr. Miranda Mafafo of statistics south Africa just to ask them one question “Now that you have found this untapped wealth of statisticians in Africa, what are you going to do with it?” Ofcourse there are a number of indicators that can describe Africa to a certain acceptable level of confidence: poverty, disease, hunger, ignorance, poor governance, presence of inhuman rights and lots of undemocratic characteristics. Measurement of these and related indicators require valid, sufficient and properly kept quality statistics at all levels especially national and regional levels. Given that we now know there exists a gigantic untapped statistics base of human resources, how do we continue from here?

Below are some of my wild suggestions formulating the famous ten point programme:

1. Support a regular young statistician’s conference to be hosted every year by different African countries
2. Build and strengthen independent National Statistical Organizations in African member states
3. Support statistics students’ exchange within African states and developed countries
4. Develop a statistics human development fund (SHDF) for African statistics scholars
5. At continental level, develop policies that will encourage development of statistics for all member countries
6. Encourage student’s attachments within and across countries
7. Encourage well planned link between statistical methods and usage or application of statistics
8. Promote demand driven short courses for both statisticians and non statisticians
9. Support infrastructural development of statistics training institutions in Africa
10. Promote and support national statistics association across all countries in Africa and their membership

I guess we can really do all this concurrently. Remember, if we can do it, then let’s do it well and since we want to go far, let’s go together as a continent and subsequently as a global village…

Alright, for an explanation and any other accompanying idea, please you are free to contact me…

On this background, therefore, we present multinomial logistic models to monitor MDGs, taking a case study of the sixth MDG of combating major diseases, particularly focusing on HIV/AIDS in one of the remotest districts of Uganda called Kapchorwa district. The period of study was three years and the survey was based on household as a source of information. Two sub counties were used, thus Kaptanya and Binyiny sub counties because these two sub counties are where SNV has had an intervention for the three years under study beginning from the year 2005.

With only about seven years left to the target year 2015, new tools such as the multinomial logistic models developed and presented in this paper become important conjugates in determining whether countries are on target to achieve the MDGs. This then acts as a pre-cursor to what resources could be required as an input so that the set targets are met in the specified period. Click to download a presentation of the paper presented at the first Africa conference of young statisticians 2008, Pretoria South Africa. You are welcome to leave a contribution or comment…

Keywords

MDG, HIV/AIDS, Multinomial Logistic Model, Monitoring