Free Essay On Statistical Hypothesis

A good study is determined by how strong, smart and eligible the study hypotheses. In addition, the study hypotheses should link directly to the study questions and objectives. A good research hypothesis should be easily traced and tested in the final findings and analysis. In a study of obesity among the children, a number of datasets would be considered appropriate for testing of possible hypotheses. Such data sets and variables would be age, socioeconomic background of the child, family size and history, leisure and other activities that the child is involved in, a child’s esteem, and parental contribution and level of household income among others.
The family would be considered an important in testing of hypothesis possibly in two ways; 1. There is a relationship between the family history and obesity and; 2. Smaller family sizes contribute to obesity. Such hypotheses would rise when one considers family as a strong variable of interest in hypotheses testing. Age is another variable of interest in testing hypothesis on obesity amongst the children. For instance, obesity is more prevalent amongst children aged four years. To test this, a number of anthropogenic variables such as BMI would be of substantial significance. Ethnicity also stands strong as a data set in obesity. The ethnic background should be considered since ethnicity plays a great role in feeding a culture of a family.
The data computed for ethnicity, housing tenure, parents’ contribution to the child’s upbringing, family size and history, the leisure of the child and age would be analyzed quantitatively by use of statistical packages like the SPSS and EpiInfo. A multilevel imputation model would also be used in the adjustment of the missing variables. Rubin’s rules are supposed to be employed in combining of the data collected. Lastly, any change in BMI would be compared with the RCT data. NVIVO would be substantial in the qualitative data analysis, reduction, and testing. Positive tests of association would be tested through correlation tests, chi-square run tests among other ways and a given p-value (p=0.001) shall be used in the test of association.