Scribd Logo
Upload

Welcome to Scribd!

  • 4 views

    Lecture 10 ReKm Parametric Tests Z-Test T-Test

    Uploaded by

    rayta.rahman

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd

    Lecture 10 ReKm Parametric Tests Z-Test T-Test

    Uploaded by

    rayta.rahman
    4 views12 pages

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    4 views12 pages

    Lecture 10 ReKm Parametric Tests Z-Test T-Test

    Uploaded by

    rayta.rahman

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    of 12

    Statistics

    Parametric Tests

    Reatul Karim
    Assistant Professor
    Department of Pharmaceutical Sciences
    Z-test (for large samples, >30)

    To test the difference between two population means, the test statistic is

    ( x1 − x2 )
    z=
    2 2
    s1 s 2
    +
    n1 n2
    Where s1
    2 and s 2 are variances of two samples
    2

    Example:
    The haemoglobin (Hb) level of children was measured in 143 girls
    and 127 boys. The results were as follows:

    Number (n) Mean (Hb%) SD


    Girls 143 11.2 1.4
    Boys 127 11.0 1.3
    Z-test (for large samples, >30)

    The test procedure is as follows:

    1. State the null hypothesis- mean Hb is same in girls as well as in boys.


    2. Find z value.

    Here n1 = 143, x1 =11.2, s1 = 1.4


    n2 = 127, x2 = 11.0, s2 = 1.3

    (11.2 − 11.0)
    Therefore, z= = 1.22
    (1.4) 2 (1.3) 2
    +
    143 127

    Since z<1.96, p>0.05, the difference is not statistically significant, i.e.


    it could easily have occurred by chance.
    An examination evaluating cognitive knowledge in
    basic pharmacology was mailed to a random sample
    of all pharmacists in a particular country. Those
    responding were classified as either hospital or
    community pharmacists. The examination results
    were:
    Hospital Community
    Pharmacists Pharmacists
    Mean Score 82.1 79.9
    Variance 151.29 210.25
    Respondents 129 142

    Is there any significant difference between the types


    of practice based on the examination results.
    • You are given two samples. Mean weight of
    the sample 1 is 58.6 kg, n=105 and SD=8.3.
    Mean weight of the sample 2 is 56kg, n=101
    and SD =6.2. Find whether the difference is
    statistically significant or not.
    t Tests
    • t tests are based on normal distribution.
    • They are applied for judging the significance
    of difference between the means of two
    different groups or same group before and
    after treatment.

    2 types of t-test

    1. Unpaired t-test (two independent samples)


    2. Paired t-test (single sample correlated observations)
    Unpaired t test
    • A comparison between the two means of two
    independent groups is made with unpaired t test.
    • The observations in one group are independent
    of those in the second group. That is, the
    individual measures in one group are in no way
    linked with or related to any of the individual
    measures in the other group.
    • For example, if we want to compare the heights of
    two different groups, male and female.
    Unpaired t test

    Criteria
    1. Data are quantitative
    2. Normal distribution.
    3. Observations are independent of each other.
    4. Sample size is less than 30.
    Unpaired t test can be used in two types of cases:
    a. the case in which variances are equal, i.e., 12 = 22
    b. the case in which variances are not equal, i.e., 12 = 22
    Situations for Unpaired t test
    • A new drug is compared to a standard drug.
    • Comparison of two different drugs for their
    efficacy given to two different groups of rat or
    human.
    • Weight, potency, dissolution time etc of tablet
    from two different batches.
    t-test (for small samples, <30)

    Unpaired t-test (in case of equal variance )


    When population variances (though unknown) can be assumed equal then the
    appropriate test statistic is Student’s t defined by

    ( x1 − x2 )
    t=
    Where 1 1 With degree of freedom = n1+n2-2
    s +
    n1 n2

    s=
     1 1  2 2
    ( x − x ) 2
    + ( x − x ) 2

    n1 + n2 − 2

    Degree of freedom = freedom of choice


    t-test (for small samples, <30)

    Unpaired t-test (in case of unequal variance )


    Sometime it is not reasonable to assume equality of the
    population variance and hence we cannot apply normal test as
    well because of the smallness of the sample size. In such case, we
    apply t-test, defined as follows:

    ( x1 − x2 )
    t =
    2 2
    s1 s
    + 2
    n1 n2

    Where degree of freedom is


    2
     s12 s2 
    2
     
     n + n 
    d. f =  1 2 
    2 2
    ( s1 / n1 ) ( s 2 / n2 )
    +
    n1 − 1 n2 − 1
    Steps in performing unpaired t test
    • State the null hypothesis
    • Calculate mean difference and pool/both
    standard deviation.
    • Find the t value.
    • Calculate degrees of freedom.
    • Choose significance level p≤0.05
    • Then interpret the result from the table.

    You might also like