Presentation is loading. Please wait.

Presentation is loading. Please wait.

Comparing means Norhafizah Ab Manan. After class, you should Understand independent t test, paired t test and ANOVA Know how to calculate the t statistics.

Similar presentations


Presentation on theme: "Comparing means Norhafizah Ab Manan. After class, you should Understand independent t test, paired t test and ANOVA Know how to calculate the t statistics."— Presentation transcript:

1 Comparing means Norhafizah Ab Manan

2 After class, you should Understand independent t test, paired t test and ANOVA Know how to calculate the t statistics Find the t tabulated from t distribution table

3 Comparing two means How can we get a mean? What data?- categorical or numerical?

4 Independent t test Is cholesterol level differ between male and female students? What is the null hypothesis for this study? malefemale

5 Independent t test Measure – Compare two means Assumptions 1.In each group of the dependent variables, the distribution is normal 2.Random sample How to test the assumption?

6 Steps in hypothesis testing 1.Define the null and alternative hypothesis  H o =The population means in two groups are equal  H a =The population means in the groups are not equal 2.Calculate the t statistics/ t calculated 3.Compare the t statistics to the value from t-distribution 4.Interpret the results

7 Example; A researcher interested to compare cholesterol level between male and female students. There are 12 males and 7 females. The data was: GroupSample size Mean (mmol/L) SD (mmol/L) Male126.1920.3919 Female75.4140.6492

8 1. Define null and alternative hypothesis. Ho= The cholesterol means in male and female students are equal Ha= The cholesterol means in male and female students are not equal

9 2. Calculate the t statistics

10 GroupSample size Mean (mmol/L) Variance (mmol/L) Male126.1920.3919 Female75.4140.6492 = 0.483

11 2. Calculate the t statistics GroupSample size Mean (mmol/L) SD (mmol/L) Male126.1920.3919 Female75.4140.6492 = 0.778/0.3305=2.36

12 3. Compare the t statistics to the value from t-distribution If t calculated > t tabulated (from table)- we reject the null hypothesis If t calculated < t tabulated (from table) – we fail to reject the null hypothesis

13 One tailed Right-tailed Sign of Ha is > Key word: More than Left-tailed Sign of Ha is < Key word: Less than Rejection area

14 Two tailed The sign of H A is ≠ Key word: no different Rejection area

15 3. Compare the t statistics to the value from t-distribution Find the t tabulated from t distribution table =2.45 (from table) with alpha error= 95%, Upper tailed = 2.5%. Degree of freedom= the smaller of (n 1 -1) or (n 2 -1) 6? T statistics=2.36 2,45 -2,45 0

16 4. Interpret the results The t calculated value is in the critical region Reject the null hypothesis There is different of cholesterol between gender

17 Paired t test Measure – Compare two dependent means (before and after) Assumptions 1.Distribution of the different is normal 2.Random sample How to test the assumption?

18 Example A researcher interested to determine the effectiveness of an intervention towards BP. The BP of the subjects were measured twice; before and after the intervention. MeasureSample size Mean of dSdSd Pre-interV156.48.48

19 1. Define null and alternative hypothesis. Ho= there is no different of BP before and after the intervention Ha= there is a different before and after the intervention

20 2. Calculate the t statistics The formula for t statistics: t=test statistics ḋ = mean of the difference S d =Sd of the difference n= sample size

21 2. Calculate the t statistics MeasureSample size Mean of dSdSd Pre-interV156.48.458 =2.930 Note: A hypothesized mean difference ( μ d ) can be any specified value. The most common value specified is zero.

22 3. Compare the t statistics to the value from t-distribution T calculated= 2.930 T tabulated with df (14) and alpha(0.05)= 2.14

23 4. Interpret the results The t calculated value is in the critical region Reject the null hypothesis There is a different between before and after the intervention

24 ANOVA To compare means between more than two groups Variable: – Independent variable: Categorical – Dependent variable: Numerical Assumptions: – Data is normal distributed – Equal variance

25 Examples To determine whether BMI is different between age groups or not To study the effect of 3 different types of anti hypertensive drug on 120 patients. To compare the mean different of IQ scores among 3 classes

26 References Basic Biostatistics statistics for public health practice. 2008. B Burt Genstman. Jones and Batlett Publisher Inc. Medical statistics at a glance. 3 rd Edition. 2009. Aviva Petrie & Caroline Sabin. Wiley-Blackwell.


Download ppt "Comparing means Norhafizah Ab Manan. After class, you should Understand independent t test, paired t test and ANOVA Know how to calculate the t statistics."

Similar presentations


Ads by Google