Scribd Logo
Upload

Welcome to Scribd!

  • 9 views

    Stat A4

    Uploaded by

    b
    This document provides an overview of statistical tests and procedures for summarizing and analyzing data in STATA. It includes instructions for creating frequency tables, graphs, descriptive statistics, histograms, boxplots, scatterplots, z-tests, t-tests, chi-squared tests, tests of normality, equality of variance tests, effect sizes, and confidence intervals. Key tests covered are one-sample tests, two-sample independent and paired tests, tests of independence, correlation, normality tests, and equality of variance tests. Steps outlined include identifying variable types, selecting the appropriate test, specifying hypotheses, and checking assumptions before interpreting results.

    Copyright:

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

    Stat A4

    Uploaded by

    b
    9 views2 pages
    This document provides an overview of statistical tests and procedures for summarizing and analyzing data in STATA. It includes instructions for creating frequency tables, graphs, descriptive statistics, histograms, boxplots, scatterplots, z-tests, t-tests, chi-squared tests, tests of normality, equality of variance tests, effect sizes, and confidence intervals. Key tests covered are one-sample tests, two-sample independent and paired tests, tests of independence, correlation, normality tests, and equality of variance tests. Steps outlined include identifying variable types, selecting the appropriate test, specifying hypotheses, and checking assumptions before interpreting results.

    Original Description:

    a

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    This document provides an overview of statistical tests and procedures for summarizing and analyzing data in STATA. It includes instructions for creating frequency tables, graphs, descriptive statistics, histograms, boxplots, scatterplots, z-tests, t-tests, chi-squared tests, tests of normality, equality of variance tests, effect sizes, and confidence intervals. Key tests covered are one-sample tests, two-sample independent and paired tests, tests of independence, correlation, normality tests, and equality of variance tests. Steps outlined include identifying variable types, selecting the appropriate test, specifying hypotheses, and checking assumptions before interpreting results.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    9 views2 pages

    Stat A4

    Uploaded by

    b
    This document provides an overview of statistical tests and procedures for summarizing and analyzing data in STATA. It includes instructions for creating frequency tables, graphs, descriptive statistics, histograms, boxplots, scatterplots, z-tests, t-tests, chi-squared tests, tests of normality, equality of variance tests, effect sizes, and confidence intervals. Key tests covered are one-sample tests, two-sample independent and paired tests, tests of independence, correlation, normality tests, and equality of variance tests. Steps outlined include identifying variable types, selecting the appropriate test, specifying hypotheses, and checking assumptions before interpreting results.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    of 2

    STAT1103 Finals — Ygnacio 4745379

    SUMMARISING DATA

    Categ uni/
    bivariate
    Create frequency table tab1 var1 OR tabulate var1 var2

    Create bar graph graph bar(count/percent), over(var1) OR


    over(var2)
    Create pie chart graph pie, over(var1)

    Numeric
    uni/bivariate
    Descriptive stats summary summarise var1, detail OR tabstat var1,
    statistics (n mean sd median iqr range)

    Create histogram histogram var1 OR histogram var1,


    by(category) freq ex. Gender
    Create boxplot one variable graph box var1

    Comparative boxplot graph box var2, over(var1) ******[var2 is


    numerical, var1 categorical]
    Create scatterplot Scatter var1 var 2 OR graph matrix var1
    var2 var3
    scatter graph twoway (scatter y x) (l t score age)
    *l t for trend line
    jitter (un-overlay points in jitter(number) ex. 7
    scatterplot)

    ONE-SAMPLE TESTS

    Z-test ztest var1 == pop mean, sd(given standard dev)


    z-score (Observed score - mean)/sd

    T-test ttest var1 == pop mean

    x2 gof [to get chi2 test statistic] csgof var1, expperc(exp freq1, expfreq2, H0: observed distrib
    expfreq3) same as expected
    distrib
    [to display] display chi2tail(df,test statistic) df = categs-1

    TWO-SAMPLE TESTS

    Independent Two-sample ttest var1, by(category)


    T-test
    Correlation test pwcorr var1 var2, sig
    Chi-squared test of tabulate var1 var2, chi2 row Testing association: are proportions of
    independence expected this variable di among values of other
    variable? 0: association, 1: no assoc
    chi2tail(df, test statistic) df = (rows-1 x
    (cols -1)
    expected frequency: (row total x
    column total) / grand total
    McNemar’s test mcci tablecontent1 2 3 4 Testing related groups categorical

    ASSUMPTION TESTS

    Shapiro-Walk normality swilk var1 OR by categoricalvar, sort: swilk


    test numericvar1
    NORMALLY distrib if p>0.05
    fi
    ff
    fi
    Levene’s test robvar var1, by(category) &W

    EQUAL var if p>0.05

    EFFECT SIZES

    Cohen’s d tests related to means


    1sample t (M-μ) / sd (sample mean - hypothesised popu mean /
    sample standard dev)
    2sample t (x̄ 1-x̄ 2) / sp (mean 1 - mean 2 / pooled standard dev)
    Paired t Md/sd (sample mean di erence / standard dev of
    di erences)
    Cohen’s W √(x2/N) (√(chi test statistic / sample size))

    CONFIDENCE INTERVALS

    Critical value: value of z/t that cuts o 5% in both tails of the normal distrib

    P-value sig: interval EXcludes mew || P-value not sig: interval INcludes mew

    General form for con dence interval: Sample estimate +/- critical value x standard error

    M +/- critical value x (variance/√n)

    If p LESS OR EQUAL </= 0.05 con dence interval


    If p GREATER THAN > 0.05 con dence interval,

    • REJECT null, YES di erence, SIGNIFICANT


    • FAIL to reject null, NO di erence, NOT signi cant

    • swilk NOT normal distributed


    • swilk NORMAL distrib

    • Robvar NOT EQUAL vars • robvar EQUAL vars

    Steps

    1. How many variables?

    2. What kind of variables? categorical/numerical?

    3. What test appropriate?

    4. What hypos in this test?

    5. Normally distributed? swilk

    6. Equal variance? Lev’s

    7. Reject/fail H0?

    8. Signi cant di ?

    Summary

    Looking at di erences:

    T-tests (numerical) Chi-square tests


    (categorical)
    Di bw known value One-sample Goodness of t
    Independent groups Independent two-sample Test of independence
    Related groups ex. Siblings, Paired t-test McNemar’s
    same people measured twice; usually
    speci ed

    Looking at similarities: correlation

    ff
    ff
    fi
    fi
    ff
    fi
    ff
    ff
    ff
    fi
    ff
    fi
    fi
    fi
    ff

    You might also like