Variables: Possible nominal variables in my research area of interest include storage of explosives, volume of financial transactions, terrorist publicity materials, rehearsals of an attack and incidences of terrorist detection by law enforcement. In this research my focus will be on studying how law enforcement can be better able to identify persons that will follow through on a terror attack; a possible nominal variable is incidences of explosive storage by peculiar individuals. This variable can be counted and related to identification of terror attacks. Another nominal variable is detection or identification of terrorists. Ordinal variables within the research will be ‘nature of financial transactions’; this will be ranked by the nature of volumes transferred between suspicious individuals and can be low, moderate or very high. Others include degree of appearance of publicity material in various locations and prevalence of rehearsals. Interval scale measures and ratio scale measures in this research are difficult to find because of the nature of terrorism i.e. it is very unpredictable and patterns cannot be generalized. (Field, 2009)
Hypotheses: There is a positive correlation between incidences of terror detection and the volume of financial transactions made by terror groups. This is the hypothesis: there is no relationship between incidences of terror attacks and volumes of financial transactions enacted by suspicious groups.
Experiment vs correlational: My area of research is more conducive for correlational rather than experimental research because experimental research would entail manipulation of variables. In this case, it is difficult, unethical and sometimes impossible to: manipulate the transfer of finance between terror groups, store explosives by these groups and public terror attacks. Manipulating these variables merely for research would cause a huge scare in the public domain and no empirical researcher would seriously try to cause such a stir. Instead, correlational research would be more appropriate because this would simply involve looking at previous incidences of terrorism detection and their relation to the other variables such as volume of finance, publicity materials and the like.
Reliability versus validity: Reliability can best be understand as the degree to which a measure is consistent. In other words, a researcher needs to ask whether he/ she can get the same results using the same procedure, sample and methodology. In my area of research which is identification of terror attacks by law enforcement, it is critical to draw the right inferences by basing them on consistent results. However, validity would be more important because it shows that the researcher is actually measuring what he ought to be measuring. In other words, it would be pointless to produce consistent results (possess high reliability) if the correlational variables were not related to one another or were not causing any outcomes (low validity).
Sample versus population: In my area of research, populations would be represented by all the possible set of objects under analysis. In this case, it would be all terrorist incidences or all volumes of finance transferred. It is important to understand these differences in a statistics course because researchers often do not have the time and resources to collect data from all members of the selected population. Furthermore, inferences made are only as accurate as the sample chosen. (Field, 2009)
Central tendency: Mean =127.1053, Mode=115, Median: 115. The data is best described by the mean because it includes all the values of the data set in its calculations, the other two measures are not indicative of how divergent the data is.
Measures of dispersion:
- Variance={∑ fx2/∑f}-{∑fx/∑f}2={348947/19}-{2415/19}2=18365.6316-16155.7526=2209.8790
- Standard deviation=√variance=√2209.8790=47.009
- Range=xmax-xmin=275-100=175
- Interquartile range=3(n+1)/4-(n+1)/4=15th value-5th value=117-105=12
Descriptive statistics: Basic descriptive statistics are important because they provide a platform against which all the data collected are first presented in manageable forms. A description of research findings is always a first step to the process of using those findings for larger generalizations. After organizing data, then one can start looking for patterns or inferences from that material. In other words, descriptive statistics is a prerequisite to inferential statistics. (Field, 2009)
Statistical significance: In the hypothesis in five i.e. ‘ there is a positive correlation between incidences of terrorism detection and volume of financial transactions made by terror groups’, a statistical significant finding would be one that obtains a critical value that is greater than the decided error rate i.e. p>alpha such that the statistical correlation between incidences of terrorism and volume of financial transactions is greater than the error rate which will have been selected through an effect size analysis or an examination of the practicality of the difference found. This would be a rejection of the null hypothesis.
Type 1 and type 2 error: A type one error in my error of research would be failing to acknowledge a null hypothesis by asserting that there is a positive association between incidences of terrorism detection and volume of financial transactions among terror groups. In other words a relationship has been established yet there is none. This would lead law enforcers to pursue the wrong channels of terror detection and would waste time and resource. A type 2 error would be acknowledging a null hypothesis. In my research I would be saying that there is no relation between incidences of terrorism detection and volume of financial transactions between terror groups. This error would result in a large loophole for terrorist attacks since law enforcers would not bother monitoring such a channel.
Reference
Field, A. (2009) Discovering statistics using SPSS (3rd ed.). Los Angeles: Sage.