Are Civil Rights and Political Freedoms correlated in NationStates?

From the archives 11.5 years ago:

Are Civil Rights and Political Freedoms correlated in NationStates?

Note: Using a simple correlational analysis on two variables only tested, this piece, will no doubt, be concise and simple unlike most correlational published studies and its reporting of results similar. There may be a chance, however, if time permits, to complete a follow-up and give an in-depth design, data mining and results. In the mean-time, I will only be testing two variables on the data points I have specified from the onset. Of course, as the raw data comes from NationStates systems, there is only so much data which can be used usefully. A main issue with secondary data such as this, I am afraid may not be overcome at ease. And this must be noted at the outset. Although, NationStates can be filled with a treasure trove of data, in particular from the sub-categories; there will always be issues in conducting any in-depth analyses.
This piece was written quickly and without much reference point: out of a curiosity with the relationship between Civil Rights and Political Freedoms, with an initial debate on the relationship and the interpretation of the data from an eye view between Philosophy 115 (P115). Any and all questions are welcome on the P115 Regional Message Board (RMB), as well as data, plots, graphs and analysis etc available on request.

The NationStates system has developed three main variables on nation classification: economy, political freedoms, and civil rights. Although all three may be useful to look at in the round, and to analyse, it is particularly the political freedoms and civil rights variables which are of interest. It has been noted in the past that the economy variable has the most fluid to movements. The economy of a nation is open to a greater fluctuation rate relative to the political and civil. A canny nation holder can max out his/her economy and trend the variable for as long as needed. However, the same cannot be said for civil rights and political freedoms.
Often, and contrary to 'common sense', the relationship between the two variables: civil rights and political freedoms, can at times appear to be at odds from a glance at the trending graphs; however this is contested and certainly does not happen all of the time, but a trend which has been of interest: the 'at odds' between the two variables. To test such an hypothesis, a correlational design was employed testing the civil rights and political freedoms variables. According to the graph data, at an "eyes view", the hypothesis that as civil rights increases, political freedoms decrease, and vice versa. This appears to however be contradictory, and warrants exploration on a representative sample.

Method

Design

The sample included the potential for all nations to be included by sampling from the World Census Report data produced daily by the Nationstates systems. On the day of data collection, the category reported was 'Stupidity'. The sample was selected by dominations of 226. That is to say, on the current population of 145143 of the time this running (08/04/2015), a nation was chosen on every 226 pages. The nation selected presented itself first on the list, or second and so on until a participant nation was selected in accordance to the criteria set out as follows.
System reported data was collected using end point data in the "trend" graph of each nation on civil rights and political freedoms.
Sample
Test the hypothesis, an initial sample size was calculated using G-Power (Faul, et al. 2009) setting the parameters in accordance to an exploratory study with no work completed previously: that is setting the effect size to medium (0.3); alpha to 0.05, and the power to 0.8. Family of tests selected was T-test; using the statistical test correlation: point biserial model (a priori).
In use of this method, a sample size of 64 was surveyed to test for statistical significance.
Criteria of participants (exclusion / inclusion criteria): to run the this effectively, nations were selected on the basis of; A) population (at least one billion population (with 10% leeway under the billion mark)) to ensure rich enough data flow) and actively answering the NationStates issues (being active within 10 days. B) have shown to have answered at least one issue within that time, again ensuring the issue data was rich enough. C) Nations recently 're-founded' were also excluded from the sampling, as well as 'erratic' login patterns - in other words, nations with recent login activity dating back 30 days or more. D) Nations with known multi / switch accounts were manually excluded from the data set which met the following: i) numbered same-named nations ii) nations with prefix which matched other nations, e.g. "The Queen's State of X"; which may have a tendency to experiment with issue data, skewing the final analyses.
The exclusion criteria is intended to remove "flat line" data sets from entering into the final analyses, as well sampling from active nations answering the system's issues giving a fair a representative account of active NationStates nations.

Results

The data sets were analysed by first testing for outliers by using one of the more strict methods do to so (Tukey, 1977) as to assess the specific correlate analysis to run. It was concluded that Pearson's correlate from the initial outlier test (that is there were no outliers from 1.5 SD from the M) was best placed.
The results suggest that political freedoms and civil rights had a strong statistically significant linear positive correlation (r = .688, n = 64, p = <.01).

Discussion

The results suggest then, that political freedoms and civil rights are linear, and when both are affected in the rankings, both have a strong tendency to follow suit of each other.
This is somewhat counter to what I had anticipated when eyeing the data graphs.
However, it must be noted that the design does not take into account several factors which have an impact, and may be useful to study. One such area would be to collect data surrounding the economic variable and factoring into a correlation table. Due to time commitments, however, this was not noted.
Another worthwhile area of exploration would be sub-categorise the data sets in accordance with the 27 nation categories, however this would be a rather large undertaken and would change the types of analyses and design fundamentally. Such an undertaking would run into serious time commitments in data collection and analyses interpretation as well as specialised computation.
Future design focus may wish to develop the inclusion criteria. The current criteria, although ensured an active representative sample of NationSates may unduly exclude the majority of initial nation data from a research point of view. As NationStates is a game with many broadly 'inactive' nations (that is to say, falls outside my current criteria). Perhaps, sampling the NationStates with less stringent inclusion criteria and agreeing on an approach on the 'next best candidate' technique may cut down sampling time from a perspective follow-up.
However, major consideration must be given using this approach as to ensure practical data is only used and not sampling from 'dead' or 'flat line data' flows which NationStates suffers greatly from. As well as sampling in a consistent method design.

References

Faul, F., Erdfelder, E., Buchner, A., & Lang, A.-G. (2009). Statistical power analyses using G*Power 3.1: Tests for correlation and regression analyses. Behavior Research Methods, 41, 1149-1160
Tukey, J.W. (1977). Exploratory Data Analysis. Reading, MA: Addison-Wesley.