How to Design a Research Study

Research Methodology 

Introduction 

In this chapter, an in-depth discussion of the research design, data collection procedures, and data analyses will be entailed. 

Research Design and Rationale 

The research design will assume a quantitative approach due to the nature of the data of interest. Quantitative research entails a variety of methods, all geared towards the systematic assessment of social phenomena, using numerical data (Watson, 2015). With this in mind, juvenile arrest and recidivism rate falls under the blanket of social phenomena. Watson (2015) further goes ahead and suggests that quantitative research is primarily used to analyze data for trends and relationships. In this study, the two hypotheses under review aim to assess the relationship between the rate of arrest and recidivism rates among juveniles in Fulton County, Georgia. Therefore, a quantitative study will be useful in meeting the objectives of the study. 

A one-time cross-sectional survey involving convicted juveniles in Fulton County, Georgia, will be used to meet the quantitative aspect of the study. Cross-sectional studies are relatively easy to carry out since they are only done once. The situation is different when it comes to longitudinal studies that are more complex and require to be conducted over several years. Aldridge and Levine (2001) note that in longitudinal studies, the main limitation is attrition, which is not present in a cross-sectional design; thus, data collected using the latter design can be generalized across a population. The study will also assume a comparative approach to determine any statistically significant difference in the rate of arrest and recidivism rates among juveniles within Fulton County, Georgia. 

Population and Sample Size 

Turner (2016) identifies the causes of juvenile recidivism as inefficiencies in the program, which are often associated with program implementation, inattention of the programs towards the risk factors of recidivism, and the actual type of the program implemented. All of these factors play an important role in this study’s findings. The study population will consist of juveniles with a conviction record irrespective of their sex, who have been admitted into a juvenile program in Fulton County. Carter (2019) states that about 60% of all juvenile convicts are aged between 14 and 17 years old. In Fulton County, 16-year-old juveniles account for about 30% of total convictions. This is then followed by juveniles aged 15 years, accounting for an additional 16%. With all this data, the study population will be narrowed down to convicted juveniles aged between 14 years and 17 years. 

The Fulton County Juvenile Court enlists numbers for all juvenile cases that were either put on probation or diverted between 2013 and 2017. The total juveniles put on probation were 5846, with 1162 being adjudicated, while those diverted were 6298, with 525 being adjudicated. Both of these numbers account for the total juvenile recidivism rates in Fulton County, Georgia. Based on Carter, the target population size for the study will be about 1012 juveniles, which accounts for about 60% of all cases. This number falls in line with the intended population size of juveniles aged between 14 and 17 years (Carter, 2019). 

The sampling procedure will use a stratified probability sampling method, whereby, the population will first be segregated based on sex, i.e., male or female. From the two groups, samples will be obtained at a ratio of 3:2, with the male population being higher since males are more likely to be adjudicated (Carter, 2019). Purposive sampling will then be used to identify the sample from the population that meets the study’s requirements, i.e., aged between 14 and 17 years, convicted of a crime but enrolled in a juvenile program, for instance, probation. Data will be gathered at the Fulton County Juvenile Court with the help of the Court clerk after obtaining permission from the Georgia District Attorney’s Office. 

The sample size will be calculated using the Fischer’s formula recommended by Mugenda and Mugenda (2003) below. 

n = 

Where: 

n = the desired sample size 

Z = the standard normal deviate at the required confidence level 

p = the proportion in the target population estimated to have characteristic being measured 

q = 1-p 

d = the level of statistical significance set 

The proportion in the target population estimated to have the characteristic being measured could not be identified. Mugenda and Mugenda (2003) recommend the adoption of a 50% estimate in such a scenario where the characteristic prevalence is unknown. 

Therefore, n = (1.96)2 × (0.5) × (0.5)/ (0.05)2 

n = 384 convicted juveniles 

Of the 384 juveniles of interest, about 230 will represent the male population, and 154 will represent the female population, as based on gender as a factor affecting the occurrence of a crime. 

Statistical Test 

In any study, this is a vital part that either makes or breaks the study. This section will entail the analysis procedures that will be implemented in the assortment of the data collected. Nayak and Hazra (2011) acknowledge the use of analysis of variance (ANOVA) when comparing more than two sets of numerical data. Chi-Square Test is primarily used to compare categorical variables. In this study, the Chi-Square Test will be used as the first statistical test to contrast the relationship between juvenile recidivism rates and juvenile programs in place to prevent this from happening, such as probation. Getting a p-value of less than 0.05 will show that there is a significant relationship between the two, after which the analysis of variance will be used to test for individual variables. 

As initially stated, ANOVA is used when comparing more than two sets of numerical data. ANOVA exists in two forms, i.e., one-way ANOVA used to compare the difference between three or more groups of a single independent variable, and MANOVA used to test how one or more independent variables affect two or more dependent variables. Due to the nature of the variables of interest in the study, MANOVA will be used to contrast between the two independent variables, i.e., juvenile arrest records and recidivism rates against dependent variables such as types of crimes being committed, the juvenile programs a convict is enrolled to among other significant factors. 

Nayak and Hazra (2011) suggest the use of linear regression to assess the association between variables. The authors suggest that an inverse correlation between the two variables will yield a negative coefficient. Also, all correlation coefficients range from 0 to 1, with 0 being the least correlated and 1 being a perfect correlation. Pearson’s correlation coefficient will be used to assess the degree at which each independent variable affects a given dependent variable. It is from all these analyses that an informed conclusion can be made, and generalized recommendations can also be forwarded to the juvenile programs. All analyses will be conducted using the Statistical Package for the Social Sciences software that will be installed on a computer. 

Assurance of Validity 

Validity is the extent to which a tool measures what it is supposed to measure. Validity in research entails a similar concept and affects the degree to which the results are accurate. In a quantitative study, validity is the extent to which any instrument measures what it is intended to measure. Validity in research entails two essential components: credibility and transferability. Credibility refers to whether the results are legitimate based on the sampling design and analyses carried out. To some extent, credibility also affects the replication of a study (Mohajan, 2017). In the case of transferability, it shows whether the results can be generalized to a similar group. 

With all these cautions in mind, the study will meet its credibility since it has used a globally accepted sampling design. The Fulton County Juvenile Court has published the numbers estimated during the population determination, thus are credible and factual. Furthermore, based on the nature of the study, the independent variables will be interrelated with the covariables to acquire informed conclusions. This process will involve a series of steps all dependent on the previous one, which is bound to make the data more and more legitimate. Additionally, the analysis procedures discussed are statistically accepted and globally used; hence, thresholds are well understood when making conclusions, for instance, the p-value in the ANOVA test. 

The transferability of the study has been achieved by obtaining a close representation of the population by choosing a sampling technique that upholds this factor. The ratio of male to female has also been catered for to ensure as close as an accurate representation of the population. Also, a description of the age bracket of concern has been detailed in the population and sample size category. This ensures that the findings of this study can be generalized across a similarly aged bracket with similar characteristics. All these are considerations that have been made to ensure the study is valid. 

Measurement of Validity 

In this case, Cronbach’s alpha, which is a general form of Kuder-Richardson (K-R) 20 formula, will be used. A value of 0.7 and above is considered acceptable, with 0.9 and above being excellent (Cronbach & Meehl, 1955). This will be done together with the analyses and reported in chapter four of this study. 

Population and Population Size 

The population of interest is juvenile convicts that have been enrolled in a program that curbs recidivism and are between the age of 14 and 17 years old. Based on the data reported by Fulton County Juvenile Court, an estimate of about 1012 juveniles meet the above interest criteria, and this is the population from which the sample will be drawn. 

Summary and Transition 

Due to the nature of the study, a one-time cross-sectional study design will be used to obtain the quantitative data needed to perform the study. From the large population size of juvenile convicts in Fulton County, Georgia, a sample will be drawn using stratified probability sampling, where the population will be segregated based on sex, i.e., male and female. This will then be followed by purposive sampling to obtain a sample size composed of both genders at the ratio of 3:2 with males being the higher proportion. This is because, based on the literature review, the probability of males being adjudicated is higher than that of females; hence, this consideration had to be made to ensure the validity of the results. 

Other than the descriptive analyses which are usually familiar to most people, inferential statistics using Statistical Package for the Social Sciences software will be conducted. The research plans to use Chi-Square and ANOVA tests to determine the significance of the relationship between the independent and dependent variables. These analyses will then be supplemented by the Pearson correlation coefficient, which will determine the nature and degree of correlation between the variables. 

Once the study has been done, chapter 4 will report on the results and findings. The findings will be contrasted against similar studies that were done in this area to identify any discrepancies among the studies. It is from this chapter that informed conclusions can be made. 

References 

Aldridge, A & Levine, K. (2001). Surveying the Social World: Principles and Practice in Survey Research. Open University Press, Buckingham . 

Carter, J. (2019). Analysis of juvenile court data for selected metropolitan Atlanta counties . Retrieved from http://www.gahsc.org/jcarter/safe/prelimresults.htm 

Cronbach, L. J., & Meehl, P. E. (1955). Construct validity in psychological tests. Psychological bulletin , 52 (4), 281. 

Mohajan, H. K. (2017). Two criteria for good measurements in research: Validity and reliability. Annals of Spiru Haret University. Economic Series , 17 (4), 59-82. 

Mugenda, O. M. & Mugenda, G.A. (2003). Research Methods, Quantitative and Qualitative approach . Acts press. 

Nayak, B & Hazra, A. (2011). How to choose the right statistical test? Indian Journal of Ophthalmology, 59 (2): 85-86 . 

Turner, J. (2016). Implementation of Behavioral Programs in Juvenile Facilities and the Impact on Juvenile Recidivism: A Review of the Literature. UC Merced Undergraduate Research Journal , 8 (2). 

Watson, R. (2015). Quantitative research. Nursing Standard (2014+) , 29 (31), 44.