basic concepts in biostatistics

Upload: monktheop1155

Post on 14-Apr-2018

223 views

Category:

Documents


0 download

TRANSCRIPT

  • 7/29/2019 Basic Concepts in Biostatistics

    1/22

    BASIC CONCEPTS IN

    BIOSTATISTICS

    Dr Peter Olutunde Onifade

    FMCPsych

  • 7/29/2019 Basic Concepts in Biostatistics

    2/22

    Definition

    Numerous definitions by numerousauthors

    Example: it is a set of concepts, rules, andprocedures that help us to:

    organize numerical information in the form of

    tables, graphs, and charts; understand statistical techniques underlying

    decisions that affect our lives and well-being;and

    make informed decisions.

    Dr P.O Onifade: Use of SPSS 2

  • 7/29/2019 Basic Concepts in Biostatistics

    3/22

    Our working definition

    It is the scientific study of numerical databased on natural phenomena.

    Scientific.. Dataquantities of information..groups of

    individuals.

    Numerical: quantified in one way or anothers Natural phenomena: but natural and

    introduced

    Dr P.O Onifade: Use of SPSS 3

  • 7/29/2019 Basic Concepts in Biostatistics

    4/22

    statistics (singular) vs statistics

    (plural) The word "statistics" is also used in

    another, though related, way. It can be

    the plural of the noun statistic, whichrefers to any one of many computed orestimated statistical quantities, such as

    the mean, the standard deviation, or thecorrelation coefficient. Each one of theseis a statistic.

    Dr P.O Onifade: Use of SPSS 4

  • 7/29/2019 Basic Concepts in Biostatistics

    5/22

    Variable

    Variable in general sense property of an object or event that can takeon different values.

    More than one variable can be measured on

    each smallest sampling unit.

    Variable in strict sense It as a properly with respect to which

    individuals in a sample differ in someascertainableA variable is measured throughcounting (eg. weight), sorting (eg. gender) orordering (severity of depression)

    Dr P.O Onifade: Use of SPSS 5

  • 7/29/2019 Basic Concepts in Biostatistics

    6/22

    Variate/datum

    any particular measured instance of avariable spoken of as a variate: eg the

    measured weight of the body weight ofthis, that, or the other person; male andfemale are variates/data of the variable

    gender.

    Dr P.O Onifade: Use of SPSS 6

  • 7/29/2019 Basic Concepts in Biostatistics

    7/22

    Classification of variable

    Classified based on Limit of possible values - Discrete vs

    continuous Variable Cause-effect 1 - Independent vs dependent

    Variable

    Cause-effect 2 - Intervention vs outcomeVariable

    Qualitative vs Quantitative Variable

    Method of measuremenr - Measurement

    scale: nominal, ordinal, interval, ratioDr P.O Onifade: Use of SPSS 7

  • 7/29/2019 Basic Concepts in Biostatistics

    8/22

    Discrete vs continuous Variable

    Discrete Variable - a variable that canassume only whole number of values

    (e.g., gender (male/female), college class(freshman/sophomore/junior/senior).

    Continuous Variable - a variable that can

    take on many different values, in theory,any value between the lowest and highestpoints on the measurement scale.

    Dr P.O Onifade: Use of SPSS 8

  • 7/29/2019 Basic Concepts in Biostatistics

    9/22

    Independent vs dependent Variable

    Independent Variable - a variable that ismanipulated, measured, or selected by the

    researcher as an antecedent condition toan observed behavior.

    Dependent Variable - a variable that is not

    under the experimenter's control It is thevariable that is observed and measured inresponse to the independent variable.

    Dr P.O Onifade: Use of SPSS 9

  • 7/29/2019 Basic Concepts in Biostatistics

    10/22

    Qualitative vs Quantitative Variable

    Qualitative Variable - a variable based oncategorical data. Nominal:

    Ordinal:

    Quantitative Variable - a variable based onquantitative data.

    Interval: meaningful distance, but no absolute

    zero Discrete no decimal continous

    Ratio e.g weightDr P.O Onifade: Use of SPSS 10

  • 7/29/2019 Basic Concepts in Biostatistics

    11/22

    The Population in biostatistics

    The biological definition refers to all theindividuals of a given species (perhaps ofa given life-history stage or sex) found ina circumscribed area at a given time.

    In statistics, population always means thetotality of individual observations (and attimes totality of individuals) about which

    inferences are to be made, existinganywhere in the world or at least within adefinitely specified sampling area limited

    in space and time.Dr P.O Onifade: Use of SPSS 11

  • 7/29/2019 Basic Concepts in Biostatistics

    12/22

    Finite and infinite population

    Finite: a concrete collection of objects orcreatures, such as the tail lengths of all

    the white mice in the world, the leucocytecounts of all the Chinese men in the world

    Infinite: outcomes of experiments, such as

    all the heartbeat frequencies produced inguinea pigs by injections of adrenalin; theexperiment can be repeated an infinitenumber of times (at least in theory).

    Dr P.O Onifade: Use of SPSS 12

  • 7/29/2019 Basic Concepts in Biostatistics

    13/22

    Census, Parameter, Statistic Census: enumeration or count of every

    member of the population. (mental healthsurvey of all Nigerians

    Parameter: summary measure of theindividual observations made in census ofan entire population. E.g., average GHQscore

    Statistic: summary measure obtained froma sample. E.g., average GHQ score of5,000 Nigerians selected for the 6

    Geopolital zonesDr P.O Onifade: Use of SPSS 13

  • 7/29/2019 Basic Concepts in Biostatistics

    14/22

    Samples in biostatics

    A sample is a subset of a population. Sampling unit is a single instance of the

    sample about which observations ormeasurements are taken. The samplingunits frequently, but not necessarily, are

    also individuals in the ordinary biologicalsense. Discuss

    Dr P.O Onifade: Use of SPSS 14

  • 7/29/2019 Basic Concepts in Biostatistics

    15/22

    Sampling frame

    A complete list of all the eligible samplingunits, from which sample is drawn for the

    study.A requirement for simple random sampling

    Dr P.O Onifade: Use of SPSS 15

  • 7/29/2019 Basic Concepts in Biostatistics

    16/22

    Dr P.O Onifade: Use of SPSS 16

    TESTS OF SIGNIFICANCE 1:Conditions for parametric tests

    The data must have normal distribution Homogeneity of variance: 1. equality of variance

    of different groups; 2. the variance of onevariable should be stable at all levels of theother variable.

    Data must be Interval or ratio , not categorical

    Independence of different set of data, that issets of data are not from same sample, excepton case of repeated sample T-test

    TESTS OF SIGNIFICANCE 2: Difference in means

  • 7/29/2019 Basic Concepts in Biostatistics

    17/22

    Dr P.O Onifade: 17

    TESTS OF SIGNIFICANCE 2: Difference in means

    Analysis

    Parametric

    test

    Non-parametric

    test

    Difference between a mean value from a

    population and that of a specified value

    One sample

    T-Test

    One sample

    ??kolmogorov_Smirnov (K-S) test

    Difference between two means from

    (two) independent samples/populations:

    parametric; non-parametric

    Two-

    Independent-

    Sample T-

    Test

    Mann_Whitney U,

    ??Kolmogorov

    Smirnov Z, Moses

    Extreme

    Reactions,a nd

    Wald-Wolfowitz

    runs

    Difference between two means from

    repeated measures of same

    sample/population

    Paired

    sample T-test

    Wilcoxon Signed

    Ranks Test, Sign,

    and McNemar Two-

    Related Sample

    Tests

    TESTS OF SIGNIFICANCE 2: Difference in means

  • 7/29/2019 Basic Concepts in Biostatistics

    18/22

    Dr P.O Onifade: 18

    TESTS OF SIGNIFICANCE 2: Difference in means

    Difference between more than

    two means from independent

    samples:

    One-Way

    ANOVA

    Kruskal-Wallis

    Test

    Difference between more than

    two means from repeated

    measures: independent variable

    is categorical; dependent is

    quantitatve

    RepeatedmeasureANOVA

    Friedman,

    Kendalls,

    Cochrans

    Difference between more than

    two means from independent

    samples with two or more

    independent variables:

    ANCOVA

    Difference between more than

    two means from independent

    samples with two or more

    dependent variables:

    MANOVA

  • 7/29/2019 Basic Concepts in Biostatistics

    19/22

    Dr P.O Onifade: 19

    Analysis

    Parametri

    c test

    Non-parametric

    test

    Correlation between two sets of

    quantitative variables (assuming

    NO cause-effect relationship: both

    variables quantitative)

    Pearson

    Correlation

    Kendall's tau_b;

    Spearman's rho

    Correlation between two sets of

    quantitative variables (assuming

    NO cause-effect relationship: both

    variables quantitative) plus

    controlling for confoundingvariables

    Partial

    correlation

    coefficient

    Partial

    correlation

    coefficient

    TESTS OF SIGNIFICANCE 3: relationships

  • 7/29/2019 Basic Concepts in Biostatistics

    20/22

    Dr P.O Onifade: 20

    Analysis

    Parametri

    c test

    Non-parametric

    test

    Correlation between two sets of

    quantitative variables (assuming

    there is cause-effect relationship:

    both variables quantitative)

    Here, we talk ofdependent and

    independent variables

    Linear

    Regression

    Logistic

    Regression

    Correlation between two sets of

    quantitative variables (assuming

    there is cause-effect relationship:both variables quantitative)

    Here, we talk of one dependent

    and two or more independent

    variables

    Multiple

    Linear

    Regression

    TESTS OF SIGNIFICANCE 3: relationships

  • 7/29/2019 Basic Concepts in Biostatistics

    21/22

    Dr P.O Onifade: 21

    TESTS OF SIGNIFICANCE 4: Reconciliation

    Analysis

    Parametr

    ic test

    Non-

    parametric

    test

    Correlation between two sets of

    quantitative variables(assuming NO cause-effect

    relationship: both variables

    quantitative)

    MANOV

    A

    Several variables, need toreduce to dimensions/

    factors eg cattel 16 factors

    PartialExplorator

    y factor

    analysis

  • 7/29/2019 Basic Concepts in Biostatistics

    22/22

    Dr P.O Onifade: 22

    THANK YOU

    FOR

    LISTENING