Introduction to Biostatistics,SPSS Instructions,2004,[Larry Winner]

1、spss instructions for introduction to biostatistics larry winner department of statistics university of florida spss windows data view used to display data columns represent variables rows represent individual units or groups of units that share common values of variables variable view used to displ

2、ay information on variables in dataset type: allows for various styles of displaying label: allows for longer description of variable name values: allows for longer description of variable levels measure: allows choice of measurement scale output view displays results of analyses/graphs data entry t

3、ips i for variables that are not identifiers (such as name, county, school, etc), use numeric values for levels and use the values option in variable view to give their levels. some procedures require numeric labels for levels. spss will print the values on output for large datasets, use a spreadshe

4、et such as excel which is more flexible for data entry, and import the file into spss give descriptive label to variable names in the variable view keep in mind that columns are variables, you dont want multiple columns with the same variable data entry/analysis tips ii when re-analyzing previously

5、published data, it is often possible to have only a few outcomes (especially with categorical data), with many individuals sharing the same outcomes (as in contingency tables) for ease of data entry: create one line for each combination of factor levels create a new variable representing a count of

6、the number of individuals sharing this “outcome” when analyzing data click on: data weight cases weight cases by click on the variable representing count all subsequent analyses treat that outcome as if it occurred count times example 1.3 - grapefruit juice study crcl 38 66 74 99 80 64 80 120 to imp

7、ort an excel file, click on: file open data then change files of type to excel (.xls) to import a text or data file, click on: file open data then change files of type to text (.txt) or data (.dat) you will be prompted through a series of dialog boxes to import dataset descriptive statistics-numeric

8、 data after importing your dataset, and providing names to variables, click on: analyze descriptive statistics descriptives choose any variables to be analyzed and place them in box on right options include: n s s n yy s y n y y n i i n i i n i i :mean s.e. :variance 1 :deviation std. :sum :mean 2 1

9、 2 1 1 example 1.3 - grapefruit juice study descriptive statistics 83812062177.638.6324.401595.411 8 crcl valid n (listwise) statisticstatisticstatisticstatisticstatisticstd. errorstatisticstatistic nminimummaximumsummean std. deviation variance descriptive statistics-general data after importing yo

10、ur dataset, and providing names to variables, click on: analyze descriptive statistics frequencies choose any variables to be analyzed and place them in box on right options include (for categorical variables): frequency tables pie charts, bar charts options include (for numeric variables) frequency

11、 tables (useful for discrete data) measures of central tendency, dispersion, percentiles pie charts, histograms example 1.4 - smoking status smkstts 199037.937.937.9 106320.320.358.2 60911.611.669.8 133225.425.495.2 2534.84.8100.0 5247100.0100.0 never smoked quit 10 years ago quit 10 years ago curre

12、nt cigarette smoker other tobacco user total valid frequencypercentvalid percent cumulative percent vertical bar charts and pie charts after importing your dataset, and providing names to variables, click on: graphs bar simple (summaries for groups of cases) define bars represent n of cases (or % of

13、 cases) put the variable of interest as the category axis graphs pie (summaries for groups of cases) define slices represent n of cases (or % of cases) put the variable of interest as the define slices by example 1.5 - antibiotic study outcome 54321 count 80 60 40 20 0 5 4 3 2 1 histograms after imp

14、orting your dataset, and providing names to variables, click on: graphs histogram select variable to be plotted click on display normal curve if you want a normal curve superimposed (see chapter 3). example 1.6 - drug approval times months 120.0 110.0 100.0 90.0 80.0 70.0 60.0 50.0 40.0 30.0 20.0 10

15、.0 0.0 30 20 10 0 std. dev = 20.97 mean = 32.1 n = 175.00 side-by-side bar charts after importing your dataset, and providing names to variables, click on: graphs bar clustered (summaries for groups of cases) define bars represent n of cases (or % of cases) category axis: variable that represents gr

16、oups to be compared (independent variable) define clusters by: variable that represents outcomes of interest (dependent variable) example 1.7 - streptomycin study trt 21 count 30 20 10 0 outcome 1 2 3 4 5 6 scatterplots after importing your dataset, and providing names to variables, click on: graphs

17、 scatter simple define for y-axis, choose the dependent (response) variable for x-axis, choose the independent (explanatory) variable example 1.8 - theophylline clearance drug 3.53.02.52.01.51.0.5 thclrnce 8 7 6 5 4 3 2 1 0 scatterplots with 2 independent variables after importing your dataset, and

18、providing names to variables, click on: graphs scatter simple define for y-axis, choose the dependent variable for x-axis, choose the independent variable with the most levels for set markers by, choose the independent variable with the fewest levels example 1.8 - theophylline clearance subject 1614

19、121086420 thclrnce 8 7 6 5 4 3 2 1 0 drug tagamet pepcid placebo contingency tables for conditional probabilities after importing your dataset, and providing names to variables, click on: analyze descriptive statistics crosstabs for rows, select the variable you are conditioning on (independent vari

20、able) for columns, select the variable you are finding the conditional probability of (dependent variable) click on cells click on row percentages example 1.10 - alcohol & mortality wine * death crosstabulation 10535215512690 83.0%17.0%100.0% 52174595 87.6%12.4%100.0% 11056222913285 83.2%16.8%100.0%

21、 count % within wine count % within wine count % within wine 0 1 wine total 01 death total independent sample t-test after importing your dataset, and providing names to variables, click on: analyze compare means independent samples t-test for test variable, select the dependent (response) variable(

22、s) for grouping variable, select the independent variable. then define the names of the 2 levels to be compared (this can be used even when the full dataset has more than 2 levels for independent variable). example 3.5 - levocabastine in renal patients group statistics 6563.83172.03270.232 6499.6713

23、1.40953.647 group non-dialysis hemodialysis auc nmeanstd. deviation std. error mean independent samples test .204.661.72610.48464.1788.377-132.750261.083 .7269.353.48664.1788.377-134.613262.946 equal variances assumed equal variances not assumed auc fsig. levenes test for equality of variances tdfsi

24、g. (2-tailed) mean difference std. error differencelowerupper 95% confidence interval of the difference t-test for equality of means wilcoxon rank-sum/mann-whitney tests after importing your dataset, and providing names to variables, click on: analyze nonparametric tests 2 independent samples for te

25、st variable, select the dependent (response) variable(s) for grouping variable, select the independent variable. then define the names of the 2 levels to be compared (this can be used even when the full dataset has more than 2 levels for independent variable). click on mann-whitney u example 3.6 - l

26、evocabastine in renal patients ranks 67.5045.00 65.5033.00 12 group non-dialysis hemodialysis total auc nmean ranksum of ranks test statistics b 12.000 33.000 -.962 .336 .394 a mann-whitney u wilcoxon w z asymp. sig. (2-tailed) exact sig. 2*(1-tailed sig.) auc not corrected for ties.a. grouping vari

27、able: groupb. paired t-test after importing your dataset, and providing names to variables, click on: analyze compare means paired samples t-test for paired variables, select the two dependent (response) variables (the analysis will be based on first variable minus second variable) example 3.7 - cma

28、x in src&irc codeine paired samples statistics 217.8381379.779222.1268 138.8151359.363516.4645 src irc pair 1 meannstd. deviation std. error mean paired samples correlations 13.746.003src & ircpair 1 ncorrelationsig. paired samples test 79.02353.095914.726246.938111.1095.36612.000src - ircpair 1 mea

29、nstd. deviation std. error meanlowerupper 95% confidence interval of the difference paired differences tdfsig. (2-tailed) wilcoxon signed-rank test after importing your dataset, and providing names to variables, click on: analyze nonparametric tests 2 related samples for paired variables, select the

30、 two dependent (response) variables (be careful in determining which order the differences are being obtained, it will be clear on output) click on wilcoxon option example 3.8 - t1/2ss in src&irc codeine ranks 9a6.8962.00 3b5.3316.00 1c 13 negative ranks positive ranks ties total irc - src nmean ran

31、ksum of ranks irc srcb. irc = srcc. test statistics b -1.805a .071 z asymp. sig. (2-tailed) irc - src based on positive ranks.a. wilcoxon signed ranks testb. relative risks and odds ratios after importing your dataset, and providing names to variables, click on: analyze descriptive statistics crosst

32、abs for rows, select the independent variable for columns, select the dependent variable under statistics, click on risk under cells, click on observed and row percentages note: you will want to code the data so that the outcome present (success) category has the lower value (e.g. 1) and the outcome

33、 absent (failure) category has the higher value (e.g. 2). similar for exposure present category (e.g. 1) and exposure absent (e.g. 2). use value labels to keep output straight. example 5.1 - pamidronate study pamidrev * sklevrev crosstabulation 47149196 24.0%76.0%100.0% 74107181 40.9%59.1%100.0% 121

34、256377 32.1%67.9%100.0% count % within pamidrev count % within pamidrev count % within pamidrev pamidronate placebo pamidrev total yesno sklevrev total risk estimate .456.293.710 .587.432.795 1.2861.1131.486 377 odds ratio for pamidrev (pamidronate / placebo) for cohort sklevrev = yes for cohort skl

35、evrev = no n of valid cases valuelowerupper 95% confidence interval example 5.2 - lip cancer pipesrev * lipcrev crosstabulation 339149488 69.5%30.5%100.0% 198351549 36.1%63.9%100.0% 5375001037 51.8%48.2%100.0% count % within pipesrev count % within pipesrev count % within pipesrev yes no pipesrev to

36、tal yesno lipcrev total risk estimate 4.0333.1115.229 1.9261.6982.185 .478.412.554 1037 odds ratio for pipesrev (yes / no) for cohort lipcrev = yes for cohort lipcrev = no n of valid cases valuelowerupper 95% confidence interval fishers exact test after importing your dataset, and providing names to

37、 variables, click on: analyze descriptive statistics crosstabs for rows, select the independent variable for columns, select the dependent variable under statistics, click on chi-square under cells, click on observed and row percentages note: you will want to code the data so that the outcome presen

38、t (success) category has the lower value (e.g. 1) and the outcome absent (failure) category has the higher value (e.g. 2). similar for exposure present category (e.g. 1) and exposure absent (e.g. 2). use value labels to keep output straight. example 5.5 - antiseptic experiment trtrev * deathrev cros

39、stabulation 63440 15.0%85.0%100.0% 161935 45.7%54.3%100.0% 225375 29.3%70.7%100.0% count % within trtrev count % within trtrev count % within trtrev antiseptic control trtrev total deathno death deathrev total chi-square tests 8.495b1.004 7.0781.008 8.6871.003 .005.004 8.3821.004 75 pearson chi-squa

40、re continuity correctiona likelihood ratio fishers exact test linear-by-linear association n of valid cases valuedf asymp. sig. (2-sided) exact sig. (2-sided) exact sig. (1-sided) computed only for a 2x2 tablea. 0 cells (.0%) have expected count less than 5. the minimum expected count is 10.27. b. m

41、cnemars test after importing your dataset, and providing names to variables, click on: analyze descriptive statistics crosstabs for rows, select the outcome for condition/time 1 for columns, select the outcome for condition/time 2 under statistics, click on mcnemar under cells, click on observed and

42、 total percentages note: you will want to code the data so that the outcome present (success) category has the lower value (e.g. 1) and the outcome absent (failure) category has the higher value (e.g. 2). similar for exposure present category (e.g. 1) and exposure absent (e.g. 2). use value labels t

43、o keep output straight. example 5.6 - report of implant leak selfrev * surgrev crosstabulation 692897 41.8%17.0%58.8% 56368 3.0%38.2%41.2% 7491165 44.8%55.2%100.0% count % of total count % of total count % of total present absent selfrev total presentabsent surgrev total chi-square tests .000a 165 m

44、cnemar test n of valid cases value exact sig. (2-sided) binomial distribution used.a. p-value cochran mantel-haenszel test after importing your dataset, and providing names to variables, click on: analyze descriptive statistics crosstabs for rows, select the independent variable for columns, select

45、the dependent variable for layers, select the strata variable under statistics, click on cochrans and mantel- haenszel statistics note: you will want to code the data so that the outcome present (success) category has the lower value (e.g. 1) and the outcome absent (failure) category has the higher

46、value (e.g. 2). similar for exposure present category (e.g. 1) and exposure absent (e.g. 2). use value labels to keep output straight. example 5.7 smoking/death by age smokerev * deathrev * age crosstabulation count 6473999040637 2042013220336 8516012260973 8573289433751 3942167122065 12515456555816

47、 8552073921594 4881979020278 13434052941872 6431119711840 7661649917265 14092769629105 smoke no smoke smokerev total smoke no smoke smokerev total smoke no smoke smokerev total smoke no smoke smokerev total age 50-54 55-59 60-64 65-69 deathno death deathrev total mantel-haenszel common odds ratio es

48、timate 1.457 .377 .031 .000 1.372 1.548 .316 .437 estimate ln(estimate) std. error of ln(estimate) asymp. sig. (2-sided) lower bound upper bound common odds ratio lower bound upper bound ln(common odds ratio) asymp. 95% confidence interval the mantel-haenszel common odds ratio estimate is asymptotic

49、ally normally distributed under the common odds ratio of 1.000 assumption. so is the natural log of the estimate. chi-square test after importing your dataset, and providing names to variables, click on: analyze descriptive statistics crosstabs for rows, select the independent variable for columns,

50、select the dependent variable under statistics, click on chi-square under cells, click on observed, expected, row percentages, and adjusted standardized residuals note: large adjusted standardized residuals (in absolute value) show which cells are inconsistent with null hypothesis of independence. a

51、 common rule of thumb is seeing which if any cells have values 3 in absolute value example 5.8 - marital status & cancer marital * cancrev crosstabulation 294776 38.137.976.0 38.2%61.8%100.0% -2.32.3 116108224 112.3111.7224.0 51.8%48.2%100.0% .7-.7 6756123 61.661.4123.0 54.5%45.5%100.0% 1.1-1.1 5510

52、 5.05.010.0 50.0%50.0%100.0% .0.0 217216433 217.0216.0433.0 50.1%49.9%100.0% count expected count % within marital adjusted residual count expected count % within marital adjusted residual count expected count % within marital adjusted residual count expected count % within marital adjusted residual

53、 count expected count % within marital single married widowed div/sep marital total cancerno cancer cancrev total chi-square tests 5.530a3.137 5.5723.134 3.6311.057 433 pearson chi-square likelihood ratio linear-by-linear association n of valid cases valuedf asymp. sig. (2-sided) 1 cells (12.5%) hav

54、e expected count less than 5. the minimum expected count is 4.99. a. goodman & kruskals g / kendalls tb after importing your dataset, and providing names to variables, click on: analyze descriptive statistics crosstabs for rows, select the independent variable for columns, select the dependent varia

55、ble under statistics, click on gamma and kendalls tb examples 5.9,10 - nicotine patch/exhaustion dose * exhstn crosstabulation count 16218 16218 13417 14418 591271 1 2 3 4 dose total 12 exhstn total symmetric measures .124.1041.166.243 .269.2201.166.243 71 kendalls tau-b gamma ordinal by ordinal n o

56、f valid cases value asymp. std. error a approx. t b approx. sig. not assuming the null hypothesis.a. using the asymptotic standard error assuming the null hypothesis.b. kruskal-wallis test after importing your dataset, and providing names to variables, click on: analyze nonparametric tests k indepen

57、dent samples for test variable, select dependent variable for grouping variable, select independent variable, then define range of levels of variable (minimum and maximum) click on kruskal-wallis h example 5.11 - antibiotic delivery ranks 181285.65 181261.15 179266.14 541 delivery 1 2 3 total outcom

58、e nmean rank test statistics a,b 2.755 2 .252 chi-square df asymp. sig. outcome kruskal wallis testa. grouping variable: deliveryb. note: this statistic makes the adjustment for ties. see hollander and wolfe (1973), p. 140. cohens k after importing your dataset, and providing names to variables, cli

59、ck on: analyze descriptive statistics crosstabs for rows, select rater 1 for columns, select rater 2 under statistics, click on kappa under cells, click on total percentages to get the observed percentages in each cell (the first number under observed count in table 5.17). example 5.12 - siskel & eb

60、ert siskel * ebert crosstabulation 2481345 15.0%5.0%8.1%28.1% 8131132 5.0%8.1%6.9%20.0% 1096483 6.3%5.6%40.0%51.9% 423088160 26.3%18.8%55.0%100.0% count % of total count % of total count % of total count % of total -1 0 1 siskel total -101 ebert total symmetric measures .389.0606.731.000 160 kappame