Statistics is a supportive measurement that has an objective to evaluate and enumerate a situation to find the most probable conclusions. It is, therefore, factual to state that statistics do not always evidence situations (Bartholomew, 2004). Rather, statistics provides an overview of the situation in a probability manner.
The accuracy of collected data depends on the strategies and techniques incorporated. It is clear that inaccuracy in collected data leads to the wrong results. Data collection, therefore, requires adequate experience and commitment to prevent failure in results.
Einstein describes data collection as a hitting and missing strategy of trying to support conclusions. He suggests that what might require considerations during data collection may be missed for another that does not (Einstein, 2000). For instance, during data collection in ecological studies, the random methods involve the collection of data relying on the randomly picked areas.
This method only allows collection of data held in the area. Also, the measurement made in that area might not all be necessary or accountable in the study. For instance, in a study to measure how perform when employed by the government, effectiveness is a vital factor to consider. Measuring effectiveness is quite hard.
This, therefore, calls for other strategies that could be counted in the study. In management, where the personality of decision making is unavailable, the decisions made by managers when giving data are impartial and unreliable. These are the types of data that cannot count. In this way, the postulation made by Einstein is workable.
There are different types of scale classified as parametric or nonparametric data (Sheskin, 2007). The distributions of parametric data can be predicted easily using the parametric tools. However, nonparametric data do not assume any distribution.
Ordinal and nominal scales are used in the nonparametric while both interval and ratio scales are used in parametric statistics. In graphical representation, we use tools such as histograms, box plots, the Q-Q, and P-P plots among others. They determine the normality of data.
In the analysis of data, there are four types of scales used in the measurement. These types are nominal, ordinal, interval and ratio scales (Louis, 1980). Nominal scale measures qualitative data. The origin of the word nominal originates from the Latin word ‘nomen’ which means name. This scale represents data that have something in common but with different names.
In this case, no data points are superior to the others. For example, data with the data points as Muslims, Christians, Hindus and Pagans are nominal scale. We observe that a Muslim is neither superior nor inferior to a Christian. All data points are equal to each other. In nominal scale, items are categorized to belong in a similar category. For instance, the four constituents written above belong to religion classification.
Just like the nominal scale, ordinal scale is a scale used in qualitative data belonging to the same category (Louis, 1980). However, unlike the Nominal scale, it has an element of hierarchy and superiority (Louis, 1986).
For example, if we would consider the category of education to classify data in different levels, we might have undergraduates, graduates, masters and PhD. In this case, PhD is a higher level than the Masters level whereas the undergraduate level is a lower level.
Unlike nominal and ordinal scale, interval scale is used for quantitative data (Louis, 1980). In this case, the data points are at similar distances from one another. For example, the data points 1, 2, 3 and 4 are in interval scale. This is because the quantitative items are at an interval of one from each other.
The point zero is used as a reference point. It allows the use of negative and the positive integers. For example, we can have a temperature of -5 degree Celsius and a temperature of 5 degree Celsius.
Ratio scale, just like the interval scale, also represents quantitative data (Louis, 1980). It measures data such as the mass, weight, amount of energy, and age among others. It is possible to make comparisons on data in ratio form because the numbers are multiples of others. In this scale, the point zero has a meaning. For example, an energy value of zero means that there is no energy. We can either divide or multiply the ratio scale by a scalar.
The statement made by British prime minister that refers to statistics as a lie applies here (Tolman, 2012). Statistics do not prove whether or not the theoretical facts presented are true. Instead, statistics supports what we already know. It, thus, implies that the conclusions we make after doing a statistical research appear to be consistent with the present knowledge.
It is not what we obtain from the statistical analysis that we always aim to investigate. In most cases, researchers are unable to collect the data for the whole population and hence consider a sample of the population. It is true that data collection relies on samples.
The results retrieved from these samples determine the population properties such as mean and median. This clearly shows that assumptions made in deriving concepts for the whole population lead to wrong conclusions. This is unrealistic and supports the statement made by the prime minister.
References
Bartholomew, D. J. (2004). Measuring intelligence: facts and fallacies. Cambridge, UK: Cambridge University Press.
Einstein, A., & Cameron, W. (2010). Not Everything That Counts Can be Counted. Quote Investigator: Dedicated to the Exploring and Tracing of Quotations. Web.
Louis, N. (1980). On the scales of measurement. Irvine: School of Social Sciences, University of California.
Louis, N., & Luce, D. (1986). Measurement: The theory of numerical assignments. Psychological Bulletin, 99(2), 166-180. Web.
Sheskin, D. (2007). Handbook of parametric and nonparametric statistical procedures (4th ed.). Boca Raton: Chapman & Hall/CRC.
Tolman, R. (2012). Lies, damn lies and statistics: Management content from Western Farm Press. Western Farm Press. Web.