In a study for ankylosing spondylitis, mean bone density amongst 2 groups of 50 people each, normal & diseased respectively is to compared. Which of the following test would be appropriate for this study?
|A||Student t test|
|C||Chi square test|
Student t test as the data is quantitative & the mean values of two different groups of 50 people are to be compared.
SIGNIFICANCE TESTS: TYPES
The type of significance test used depends on whether the data is parametric (usually normally distributed) or non-parametric
Student's t-test - paired on unpaired
Pearson's - correlation
Paired data refers to data obtained from a single group of patients, e.g. measurement before and after an intervention.
Un-paired data comes from two different groups of patients, e.g. comparing response to different interventions in two groups.
Parametric tests: These are - t test, z-test and ANOV A (or F-test) These share few common features:
i. These hypothesis refers to certain population parameters which is the population mean (in case of t and z tests) or the population variance (in case of F test)
ii. These hypothesis concern interval or ratio scale data (e.g. weight, height, BP etc)
iii. They assume that the population data are normally distributed.
a. Involves the same steps as t-test
b. Used when the sample is large (n> 100). When the total no. of observations is ≤100, use T test in place of Z test.
c. T-test though is more important as there is no situation where as Z test can be used and a t-test cannot be but the vice versa can occur.
Whereas a T test is appropriate for making just one comparison (between two sample means, or between a sample mean and a hypothesized population mean), when more than one comparison is being made (i.e. when means of> 2 groups are being compared), ANOVA (analysis of variance) is the appropriate technique.
Wilcoxon matched-pairs - compares the medians obtained
Mann-Whitney - unpaired data
Spearman, Kendall rank - correlation
Chi-squared test - used to compare proportions or percentages
Chi-square test offers alternate method of testing significance of difference between two proportions.
More than 2 groups can be compared.
Test Null Hypothesis, applying Chi-square test, finding degree of freedom and probability tables are components.
When p valve is <0.05 null hypothesis is true and the study is rejected, and vice versa.