"user manual"
Contents
1. a mn 4 gt Data View A Variable View meel Starting SPSS Processor When you wish to access a still unused existing file such as the datal file found on the http www basisboekstatistiekmet spss wolters nl website you click Cancel The small welcome screen then disappears while the empty data matrix remains on screen Now click File in the menu bar on the top left corner of the screen followed by Open and Data You are then asked to indicate the name and location of the file Typing in c data data1 for example and then clicking Open causes the sample data1 data file from the website to be opened You will have needed first to have saved the data from the website to the data directory on your c drive as a file named data1 Cancel File Open Data type for example c data datal Open VV Vv VVV 1 4 How does SPSS for Windows work 31 32 The data in the data data file will then be entered in the SPSS SPSS data matrix data matrix You can see this in Figure 1 5 data1 SPSS Data Editor 18 x File Edit View Data Transform Analyze Graphs Utilities Window Help teh 3 l4 gt ele a el SENS Ol el 1 WLTHi MSTAT yes yes no no would are certainl are lis ce2. Introduction t0 stati CS V it SPSS A guide to the processing analysing afs and reporting of research c Translation of the second fully revised edition have a frequency research question Variable is nominal 5 1 1 descriptive counting frequency table percentage mode 5 1 2 graphic pie chart bar chart Variable is ordinal 5 1 1 descriptive counting table percentage mode rank order median 5 1 2 graphic histogram Variable is interval ratio 5 1 1 descriptive stem and leaf frequency table classes percentage mode median mean standard deviation variance 5 1 2 graphic histogram and standard deviation have sub groups 5 2 descriptive using all of the above mentioned techniques for each sub group 5 3 graphic clustered bar chart box plot have a comparative research question Grouping variable is nominal test variable is nominal 6 1 Contingency table Population Sample Chi square 6 1 Grouping variable is nominal test variable is ordinal 6 2 Comparing mean rank orders Population comparing mean rank orders 6 2 1 Dependent sample two Wilcoxon 6 2 3 more Friedman 6 2 4 Independent sample two Mann Whitney 6 2 1 more Kruskal Wallis 6 2 2 Grouping variable is nominal test variable is interval ratio 6 3 Comparing means Population comparing mean rank orders 6 3 1 Dependent sample two
3. The nature of the general research question will provide the basis for answering the question about which statistical technique to use The general research question always contains one or more H1 How do prepare myself to work with this Introduction level of measurement nominal level of measurement specific research questions to which an answer must be given Introduction In general there are three types of specific research questions that can be distinguished in a research inquiry 1 Questions concerning how often or to what extent something occurs frequency An example of this type is To what extent are people in the Netherlands happy or What percentage of people possess cars 2 Questions concerning a comparison Example Are men hap pier than women 3 Questions concerning a correlation Example Is there any correlation between wealth and happiness It is clear that research questions 1 to 5 Introduction can be characterized as frequency research questions Research Ques tion 1 is for example concerned with the number of people that own such consumer products as DVD players Chapter 5 presents an example of the analysis used for data involving this type of query Research questions 6 to 9 are comparative research ques tions The analysis of data involving this type of inquiry is dealt with in Chapter 6 Questions 10 and 11 are correlation research questions In Chapter 7 we examine
4. 6 3 1 6 3 1 6 3 1 SN ie 6 3 1 5 6 3 2 6 3 2 1 6 3 2 2 6 3 2 3 6 3 2 4 6 3 2 5 6 3 3 6 3 3 1 6 3 3 2 6 3 3 3 6 3 3 4 6 3 3 5 7 1 Zll 7 1 2 7 1 3 7 1 4 7 1 5 7 2 7 2 1 7 22 7 2 3 7 2 4 Comparative questions involving an interval ratio test variable and a nominal grouping variable 132 Comparative questions involving two independent samples ttest 132 What is a t test 132 When do I use a t test 133 How do I perform a t test for two independent samples 133 How should I read the results of a t test for two independent samples 136 How should I interpret and report the results of a t test for two independent samples 138 Comparative questions involving two paired samples t test for two paired samples 138 What is a t test for two paired samples 138 When do I use a t test for two paired samples 139 How do I perform a t test for two paired samples 140 How should I read the results of a t test for two paired samples 142 How should I report the results of a t test for two paired samples 142 Comparative questions involving three or more independent samples One way ANOVA 143 What is a one way analysis of variance one way ANOVA 143 When do I use a one way analysis of variance one way ANOVA 144 How do I perform a one way analysis of variance one way ANOVA 145 How should I read the results of a one way analysis of variance one way ANOVA 149 How should I report the
5. How should I interpret the output What does the output mean for my research question e How should I present my output in my research report e How do I write up my conclusions We will provide examples of SPSS output for each presented technique On the basis of these examples we will explain how the data should be read and what they mean To illustrate this lat ter point we additionally demonstrate how the output can be re ported What are the instructional principles underlying this book A lot of consideration has been given to this book s instructional methodology Each chapter starts with an introduction that in a general sense provides a glimpse of that chapter s contents Each chapter ends with a list of key words in the order that they appear in the text Additionally the designated SPSS procedures are listed on a section by section basis so that anyone who has an excellent or even reasonable command of SPSS for Windows need not work through the entire chapter again The main instructional premise is that you learn by doing It is also important that you know what you are doing For this reason we also briefly explain the essential points of the technique to be used without going into detail To gain insight into the research material we advise you first to analyze the data graphically First construct something like a scatter diagram to see what type of relationship is possibly involved Depending on the research que
6. Introduction sample inductive or inferential statistics units population study sample study To save costs you could also solicit responses from just a portion sample of the employees that you select at random from the total employee pool Of course you still would like to make state ments about the entire employee population In such a case use must be made of inductive or inferential statistics to allow you on the basis of a particular case a sample to make general state ments about the population Before you can begin the analysis of your data you must therefore pose the question concerning which units who or what are the subject of your statements When these are only the persons or items involved in your research it is then possible to speak of a population study If you would also like to make statements about persons or items not involved in the research but so to speak re presented by the research units that you have selected you are then undertaking a sample study In Section 1 4 we briefly deal with a few statistical terms that are continually encountered when testing if the results in a sample are due to chance or if within a certain margin of uncertainty it is possible to make generalisa tions concerning the population from which the sample was drawn 1 3 A few general statistical terms The purpose of descriptive statistics is to present data in a clear and well organized manner If you have
7. N 2213 In such a case use must be made of inductive or in ferential statistics The question then arises to what extent the mean IQ found to be 99 7 is representative of the total employee population In other words what is the probability of the popula tion having a mean IQ of 99 7 if you could include all of it in the research This probability is of course not so great For the mean that has been determined for the sample is somewhat de pendent on the coincidental composition of the sample group If we were to select another random sample and still another etc it is quite probable that the mean IQ could possibly be somewhat higher or lower The values will likely deviate a little from each other but probably not very much SPSS can be used to calculate this so called standard error Section 5 1 The standard error in dicates the extent to which the computed sample mean is a good estimate of the population mean The larger the difference of IQ within the group and the smaller the sample size the greater is the standard error The standard error is therefore determined by sample size and sample homogeneity On the basis of the stan dard error you can for example state with at least a 95 cer tainty that the population mean lies between the sample mean minus twice the standard error on the one hand and the same sample mean plus twice the standard error on the other The notion of certainty or probability plays an importa
8. a nominal grouping variables 116 Comparative questions involving two independent samples Mann Whitney Utest 116 What is the Mann Whitney U test 116 When do I use the Mann Whitney U test 117 How do I perform a Mann Whitney U test 117 How should I read the results of a Mann Whitney U test 119 How should I report the results of a Mann Whitney U test 120 Comparative questions involving three or more independent samples Kruskal Wallis test 121 What is the Kruskal Wallis test 121 When do I use the Kruskal Wallis test 121 How do I perform a Kruskal Wallis test 121 How should I read the results of the Kruskal Wallis test 123 How should I interpret and report the results of the Kruskal Wallis test 124 Comparative questions involving two paired samples Wilcoxon Signed ranks test 125 What is the Wilcoxon Signed ranks test 125 When do I use the Wilcoxon Signed ranks test 126 How do I perform a Wilcoxon Signed ranks test 126 How should I read the results of a Wilcoxon Signed ranks test 127 How should I report the results of a Wilcoxon Signed ranks test 128 Comparative questions involving three or more related samples Friedman test 128 What is the Friedman test 128 When do I use the Friedman test 129 How do I use SPSS to perform a Friedman test 129 How should I read the results of a Friedman test 131 How should I interpret and report the results of a Friedman test 131 15 16 6 3 6 3 1 6 3 1
9. a total score for happiness The sum which is accomplished with the help of some SPSS commands constitutes a new variable You should have noticed that three of the five statements of un hap piness have a negative formulation for example Life is diffi cult It would be helpful if all the scores for the happiness items 1 1 How do prepare the data for analysis 23 24 recode command operationalisation item analysis homogeneity pointed in the same direction In that way a high score is always an indication of a high degree of happiness Using a recode com mand you can once the data has been entered adjust the values of all the variables by means of a single action so that 5 becomes 1 etc In Section 3 1 we explain how you can do that Sometimes you need to use more than one indicator to measure a concept We have used five indicators to measure the concept wealth This translation of an abstract concept such as wealth into measurable concrete characteristics such as car ownership is called operationalisation If you use more than one indicator to measure a concept such as wealth or happiness and then wish to combine these into one total score you must invoke some crite ria to check if this procedure is allowed In short the various op erationalisations must all measure the same thing and therefore be homogeneous With the aid of the homogeneity coefficient alpha also kn
10. appears You can now do a number of things You can begin processing the Analyze data You can also perform analyses by means of Analyze or Graphs compose graphs using Graphs In subsequent chapters we will use concrete examples to elaborate these possibilities Questions that undoubtedly occur to you are How can I shut the SPSS program off again and How can I obtain help when work ing with SPSS You shut the SPSS program down by clicking on the small black cross in the top right corner of the screen When the program is shutting down you are asked if you wish to save any changes that you may have made to the file If you indicate that you will save the modified file you are requested to give the file a name If you type in a new name the modified file will be stored in addition to the unmodified version If you leave the name unaltered the modified file will be written over the earlier version of the file When working with SPSS you can receive help by clicking Help in the menu bar displayed on screen In the roll down H1 How do prepare myself to work with this Introduction menu that appears you subsequently select Topics You then encounter the help menu which is illustrated in Figure 1 6 If for example you no longer remember how you saved your data you can type in save By clicking on data files and Display you obtain information about the procedure for
11. founded decisions Good luck G J Hulzebos RIS Registered Information Scientist Country Manager Products SPSS Benelux B V Foreword What would we like to achieve in this book Once you have studied this book you should 1 Have mastered the basic skills enabling you to work with SPSS on your own This book provides practical instructions for processing and analysing research data using SPSS for Windows SPSS stands for Statistical Products and Service Solutions currently one of the most widely used software packages for entering and statistically analysing data The basic SPSS package contains programs for such commonly used statistical analysis techniques as those in volved in calculating a frequency distribution cross tabulation or correlation coefficient In this Introduction to Statistics with SPSS for Windows we hope that you will learn through experience to trust the software package After a general introduction and familiarisation with the basic principles of SPSS you will immediately be able to enter data into the computer and to analyze it For this reason we do not discuss the many extra features of SPSS in this book The pur pose of this book is to give you the basic skills so that you can work independently with SPSS You will then inevitably learn how you can enhance these skills We invite the reader to begin by experimenting with the sample database concerning the relationship between wealth and happi
12. or a histogram 85 When do I use a pie chart a bar chart or a histogram 86 How do I create a pie chart a bar chart or a histogram 86 How do I read the output of the Frequencies procedure 89 How can I present a frequency distribution 90 How should I interpret and report a frequency distribution 91 How do I calculate the frequencies of subgroups 92 How do I compare sub groups 94 What is a clustered bar chart and a box plot 94 When do I use a clustered bar chart or a box plot 95 How do I make a clustered bar chart 95 How do I construct a box plot 97 How do I make variables comparable 101 6 1 6 1 1 6 1 2 6 1 3 6 1 4 6 1 5 6 2 6 2 1 6 2 1 6 2 1 6 2 1 6 2 1 6 21 6 2 2 new 6 2 2 1 6 2 2 2 6 2 2 3 6 2 2 4 6 2 2 5 6 2 3 6 2 3 1 6 2 3 2 6 2 3 3 6 2 3 4 6 2 3 5 6 2 4 6 2 4 1 6 2 4 2 6 2 4 3 6 2 4 4 6 2 4 5 How do I analyze my data when a comparative research question is involved 103 Introduction 104 Comparative questions involving nominal test and grouping variables Contingency tables and Chi squares 106 What is a contingency table and Chi square 106 When do I use a contingency table or Chi square 108 How do I compose a contingency table and to calculate a Chi square 109 How should I read the results of a Chi square test 113 How should I report the results of a contingency table or a Chi square test 114 Comparative questions involving ordinal test variables and
13. t test 6 3 2 more MANOVA not in this book Independent sample two t test 6 3 1 more one way ANOVA 6 3 3 have a correlation research question Variables are nominal based on the variables with the lowest level of measurement 7 1 Cram r s V Determining significance when a sample is involved Variables are ordinal based on the variables with the lowest level of measurement 7 2 Spearman Determining significance when a sample is involved 7 2 Variables are interval ratio based on the variables with the lowest level of measurement Two variables disciptive Pearson 7 3 2 Determining significance when a sample is involved vaststellen 7 3 2 imple regression 7 3 3 graphic 7 3 1 scattergram regression line More then two variables predictive multiple correlation regression 7 3 4 controlling for partial correlation 7 3 4 Introduction to Statistics with SPSS Dr D B Baarda Dr M P M de Goede en Dr C J van Dijkum Second completely revised edition Noordhoff Uitgevers bv Groningen Houten Cover design Total Identity Amsterdam Text Layout Ebel Kuipers Sappemeer Any comments and remarks about these or other publications can be sent to Noordhoff BV Higher Education Department Antwoordnummer 13 9700 VB Groningen e mail info noordhoff nl 2345 13 12 11 2004 Baarda and De Goede c o Noordhoff Uitgevers bv Groningen Houten The Nether
14. I enter my data in the computer 35 Introduction 36 How do I make a coding scheme or codebook 36 How do I enter my coding scheme in the computer 38 How do I enter my data in the computer 44 What is a data matrix 44 Using Data View to enter data in the Data Editor 44 How do I save my entered data 48 How do I check if I have made any mistakes when entering my data 49 How do I retrieve my entered data 51 13 14 3 1 3 2 3 3 3 4 3 5 4 1 4 2 o im o u a w N e 5 3 5 3 5 3 5 3 5 4 1 1 1 1 1 2 1 2 1 1 2 2 1 2 3 1 2 3 4 How do modify or combine data 53 Introduction 54 How to recode data 54 How do I save the recoded data 58 How can I divide a variable into classes 58 How can I combine data 61 How can I take missing values into account when combining data 63 How do I verify the homogeneity of the composite scores 69 Introduction 70 How can I examine if a composite score or scale is homogeneous 70 How can I determine which items are good and which ones are bad 73 How do I analyze my data when a frequency research question is involved 77 Introduction 78 The making reading presenting interpreting and reporting of a frequency distribution 79 What is a frequency distribution 79 When do I use a frequency distribution 82 How do I generate a frequency distribution 83 How do I make graphic representations of frequencies 85 What is a pie chart a bar chart
15. a A data use key 25 26 ordinal level of measurement interval and ratio levels of measurement ratio level interval level Scale Nominal Ordinal continuous variables discrete variables descriptive statistics population Such gradations can indeed be made for data measured on the or dinal interval or ratio levels Data on the ordinal level of measurement can definitely be de scribed in terms of more or less but this difference between cat egories cannot be expressed as a numerical value For example level of education is clearly an item describable as a comparison some having more and others less The upper high school level is higher than the lower one but it is not possible to indicate how much higher At the interval and ratio levels of measurement the difference between the categories conceivable as more or less is also ex pressible as a number Temperature provides a good example of this type The difference between 5 and 10 degrees Celsius is just as large as the difference between 45 and 50 degrees In contrast to the interval level the ratio level involves a natural zero point such as illustrated by weight or height The interval level in volves comparable intervals but no natural zero point This dis tinction has consequences for the arithmetic calculation that may be used With temperature it is not possible to say that 20 de grees is twice as must as 10 degrees With weight it is conversel
16. ables Pearson s product moment correlation 171 What is a Pearson s product moment correlation 171 When do I calculate a Pearson s product moment correlation coefficient 172 How do I calculate a Pearson s product moment correlation coefficient 172 How do I read the results of a Pearson s product moment correlation 173 How do I make a report on a Pearson s product moment correlation procedure 173 A simple regression analysis 174 What is a simple regression analysis 174 When do I use a simple regression analysis 175 How do I calculate a simple regression 175 How do I read the results of a simple regression analysis 176 How do I present the results of a simple regression analysis 177 Correlations involving more than two variables that are measured on the interval or ratio level Multiple correlations and partial correlations 178 What is a multiple correlation and regression 178 When do I use a multiple regression analysis 179 How do I calculate a multiple correlation 179 How do I read the results of a stepwise multiple regression 181 How do I report a multiple correlation and regression 181 What is a partial correlation 182 When do I use a partial correlation 183 How do I calculate a partial correlation 183 How do I read the results of a partial correlation 184 How do I report a partial correlation 184 Appendix 1 How can include SPSS output in my text 187 Appendix 2 How can edit SPSS ou
17. an then be stated Is there a positive correlation between the amount of wealth at an individual s disposal and the degree of happiness that the individual enjoys The concept wealth is defined by the re searcher as follows wealth is the quantity of financial means at an individual s disposal This definition is deliberately broad In this way the research not only covers income and assets but also other financial resources that individuals might have available to them Happiness is defined as follows happiness is the degree to which an individual is content about the life that he she leads The researcher has operationalized these two concepts in the form of a questionnaire To measure both the concept of wealth and the concept of happiness the re searcher has devised five statements or items for each Wealth is measured by asking respsondents to agree or dis agree with the following statements 1 lownacar yes no 2 own a home apartment yes no 3 own a DVD player yes no 4 Iam covered by the Dutch national health plan yes no 5 receive a rent subsidy yes no Happiness is measured by asking repsondents to complete the following multiple choice items 1 If could live my life over again would in the same way certainly not live it to some ive it definitely not live it extent not live it ive it 2 Most other people are than am certai
18. best suits your research question You will find the diagram on the inside front cover e How you use SPSS to conduct the analysis and how you should interpret the results This we indicate for each statistical technique to be discussed Additionally we then examine how you can make a report about the results This chapter fulther deals with a number of important statistical terms such as normal distribution significance probability and standard error Section 1 3 The final section of the chapter will explain how you can start using SPSS Befor explaining the uses od SPSS we will first introduce the practice example Wealth and happiness see Figure 1 1 The questionnaire in Figure 1 1 was presented to a representative sample of 500 men and 500 women between the ages of 25 and 55 years old These age limits were deliberately chosen Many younger people are still studying and do not therefore have any fixed income For people above the age of 55 there is often the issue of their partial withdrawal from the work force as a result of which they enter another financial category The data from this research is found on the website http basisboekstatistiekmet spss wolters nl in the file entitled data1 21 Introduction 22 Can money buy happiness A researcher would like to know if there is a correlation be tween wealth and happiness He asks himself if money can buy happiness His general research question c
19. d range instead of complicated SQL queries summary tunchones transformations Data transformations Transformation features help get your data ready for analysis scale model You can easily subset data combine categories add aggregate merge split and transpose fles and mare Overview Viewer The Viewer makes it easy to browse your results selectively show and hide output change the display order results and move presentation quality tables and charts between SPSS and other applications Multidimensional pivot tables Your results come alive with multidimensional pivot tables Explore your tables by rearranging rows columns and layers Uncover important Electronic distribution Sond e mail roperts to others with the chek of a button or Finally we draw your attention to all the various tutorials includ ed in SPSS When you open SPSS you see among the series of op tions that there is one entitled Run tutorial see Figure 1 4 Clicking this option gains you access to a list of tutorials In addi tion SPSS has its own website containing all types of information about the program www spss com 1 4 How does SPSS for Windows work 33 34 Key words u u Introduction research cycle research questions 1 1 data preparation coding scheme codebook data editor variation recoding operationalisation item analysis homogeneity 1 2 1 inquiry freguency research question comparative research q
20. determined the job satis faction of nearly a thousand employees it makes little sense to present all this data It is mostly for example compiled into a his togram Section 5 1 2 or formulated as percentages or an average 5 Std Dev 16 40 z Mean 99 7 N 198 0 60 0 80 0 100 0 120 0 140 0 70 0 90 0 110 0 130 0 1Q Figure 1 2 The distribution of the intelligence level 1 3 A few general statistical terms 27 normal distribution Gauss curve inductive or inferential statistics standard error sample size homogeneity probability 28 Section 5 1 1 You describe your data in a reduced and therefore orderly form If the data are reported in a graphic manner the re sult is quite often the so called normal distribution Figure 1 2 provides a fictive example of this It is the distribution of scores from an intelligence test adminis tered to 198 employees of the Labour company This distribution closely approximates the form of a normal distribution By way of comparison the normal distribution is sketched Such a distribu tion is also named the Gauss distribution or the Gauss curve The characteristic feature of the normal distribution is its bell form With SPSS you can verify if the distribution of your data ap proaches the normal distribution Section 5 1 2 Consider the possibility that 198 employees constitute a random sample from the total employee pool of the Labour company
21. do people feel Are there differences between men and women concerning ownership of a car home or DVD player national health plan coverage reception of rent subsidies 7 Is there any difference in people s happiness about life when they are living with or without a partner with or without children H1 How do prepare myself to work with this Introduction coding scheme data editor variation 8 Are there any differences in the loneliness felt by men and women 9 Do men and women experience different degrees of happiness about the life that they are living 10Is there a correlation between contentment about life and age 11 Is there any correlation between wealth and happiness 1 1 How do prepare the data for analysis Before you enter the data you first have to make for your own use a summary of the variables involved in the research and the values that these could have For example the variable sex ap pears in our research project and has the possible values male or female It is useful to convert these values to numerical form for example 1 for male and 2 for female You must of course appropriately note which values stand for what You do this by means of a coding scheme or codebook In Section 2 1 we explain how you can construct such a scheme When you have collected the research data and composed a cod ing scheme you can enter the data in the computer You do so by t
22. ethods are systematically presented and explained on the basis of concrete examples always related to an explicitly stated specific research question The sample research questions are developed in a step by step manner and illustrated in the boxes intended for that purpose The data files on which we conduct the analyses can be found on the internet at http www basisboekstatistiekmetspss noordhoff nl Spring 2004 Ben Baarda Martijn de Goede Cor van Dijkum Required hard and software This book explains how research data can be entered prepared analyzed and interpreted using SPSS for Windows and how a re port can be written about the output You must be able to work with Windows version 98 2000 ME or XP and must of course have installed SPSS In this book we make use of SPSS Version 12 0 11 12 Symbols and data files used The following symbols are used in this book gt Move the mouse pointer to the position indicated on the screen gt Click the Plbutton in SPSS Single click the left mouse button Double click the left mouse button The following data files are used in this book datal data2 data3 data4 data5 family raw unprocessed data entered according to the scheme of codes presented in Figure 2 1 recoded items rwlth1 rwlth2 rwlth3 rhap3 rhap4 and rhap5 have been entered here see Section 3 1 the variables ac age class Section 3 3 Twlth a
23. examples of data analysis for research involving this type of question 1 2 2 At what level can measure my data Once you have determined the type of research question s in volved in your research see the first column in the block dia gram How do I analyze my data you must then decide at which level of measurement the variable s has been measured To do this see the row below frequency comparison or correlation in the above mentioned block diagram For each research question it is necessary to indicate which level of measurement is being used for the variable sex involved In Section 1 1 the level of measurement for the variable sex in Re search Question 9 on the differences in the extent to which men and women feel themselves to be happy is different from and lower than the level of measurement for the variables in Re search Question 10 for example on the correlation between age and happiness For the sex variable there are only two categories namely male or female There is only a straightforward difference involved no degree of more or less A man is different from a woman but not more or less than her The same holds true for marital status people are married living together or single This type of possi ble answer concerns a nominal level of measurement You can in dicate how many men or women own a car but not that someone is more male or more female 1 2 How do I analyse my dat
24. icking its entry in this list gt SPSS icon In SPSS Version 12 0 you then see the welcome screen as shown in Figure 1 4 This welcome screen offers you various options If you wish for example to enter new data you have to click on Type in data and then OK An empty data matrix then appears In earlier SPSS versions the first screen that you saw was the empty data matrix In Chapter 2 it is explained how you can enter data in the now open data editor You can also open a previously used existing data file If it is a file that you have previously used you will probably find it under the heading Open an existing data source You then click on the appropriate file followed by OK and the completed data matrix is displayed on screen H1 How do prepare myself to work with this ntroduction tj Untitled SPSS Data Editor File Edit View Data Transform Analyze Graphs Utilities Window Help al le SIEN l zals al 1 5 elel B SPSS 12 0 1 for Windows What would you like to do KEN Run the tutorial Type in data Run an existing query sa C Create new query using Database Wizard Open an existing data source Open another type of fle More Files T Don t show this dialog in the future oc eme
25. lands All rights reserved No part of this publication may be reproduced stored in a retrieval system or transmitted in any form or by any means elec tronic mechanical photocopying recording or otherwise without the prior written permission of the publisher Insofar as the copying of this publication is permitted by article 16B of the Copyright Act 1912 j by the Order of 20 June 1974 Bulletin of Acts and Decrees 351 as revised in the Order of 23 August 1985 Bulletins of Acts and Decrees 471 as well as by article 17 of the Copyright Act 1912 the fees due are payable to the Reproduction Rights Association Post Of fice Box 3060 3130 KB Hoofddorp The publisher must be consulted be fore one or more sections from this publication may be included in an thologies readers or other compilations article 16 Copyright Act 1912 ISBN ebook 90 018 49887 ISBN 90 207 32978 NUR 916 Preface The extent to which a decision is accepted does not merely de pend on the position in an organisation that the decision maker occupies nor on a person s years of experience in the field the degree of expertise in the area powers of intuition or the fact that an individual may have a good nose for business Such qualities are no longer sufficient even if decisions taken on such bases are nevertheless good ones After all to be accepted decisions not only have to be right they must also be justifiable Subjectivity is only one of
26. lity of a difference and certainly not about the direction of the difference your test is then two tailed The determination of significance is based on a few characteris tics of the sample in particular its size and homogeneity The larger the sample the smaller is the probability that any finding may be due to chance The smaller the differences i e range of a given variable within a group homogeneous groups the small er is the probability that the differences between groups involves a coincidence SPSS output also frequently contains the term degrees of free dom df The number of degrees of freedom indicate the extent to which scores can vary If you know only one of two numbers i e 36 and you know that the mean is 40 the other number must then be 44 This represents 1 degree of freedom If you in fact know one figure then you also know the other For many tests such as the t test Section 6 3 1 the number of degrees of free dom is equal to the number of elements in the sample minus 1 For a contingency table see Section 6 1 the number of degrees of freedom is equal to the number of rows minus 1 multiplied by the number of columns minus 1 For a 2x2 table the number of degrees of freedom therefore equals 1 If the marginals row and column totals of a 2x2 contingency table are known as well as one of the cell frequencies you can then calculate the other cell frequencies Degrees of freedom are important when you wish t
27. nd Thap respective totals of the 5 wealth and 5 happiness items see Section 3 4 have been entered here the variable mwlth the means of the 5 wealth items see Section 3 5 and the items rwlth1_1 to rwlth5_1 for the wealth items the missing values are replaced by a series average see Section 3 5 are entered here the variable Twlth and Thap are here the respective sums of rwlth2 wlth4 and wlth5 and hap1 hap3 and hap4 see Section 4 2 this is a data file containing the happiness scores of twen ty men and their wives along with the happiness scores from each of their oldest children The data files are stored on the website http www basisboeksta tistiekmetspss wolters nl It would be most helpful if these files were copied to a special data folder directory on the c drive hard drive of the computer on which you are working Contents 1 1 1 2 1 2 1 1 2 2 1 2 3 1 3 1 4 2 1 2 2 2 3 2 3 1 2 3 2 2 4 2 5 2 6 Preface 5 Foreword 7 Required hard and software 11 Symbols and data files used 12 How do I prepare myself to work with this Introduction 19 Introduction 20 How do I prepare the data for analysis 23 How do I analyze my data A data use key 24 What specific research questions are involved in my general research question 24 At what level can I measure my data 25 Is a population or asample involved 26 A few general statistical terms 27 How does SPSS for Windows work 30 How do
28. ness that we provide This database contains survey data from 500 men and 500 women We use it to investigate the usual prac tices involved in a research project e We begin with the entering of data Chapter 2 e we then explain how SPSS can be used to modify and to com bine the data Chapter 3 e and in Chapter 4 we discuss how to establish the homogeneity of composite scores After this preparatory work it is time for analysis The methods used will depend on several factors including the specific nature of the inquiry s general research question A well designed gen eral research question always involves one or more specific is sues to which answers are sought by means of research These specific research questions can further be distinguished into e frequency research questions which ask how often or to what extent something occurs techniques for describing such vari ables are investigated in Chapter 5 e comparative research questions which establish and test the similarities and differences between two or more groups con cerning a given characteristic they will be discussed in Chap ter 6 e correlation research questions which establish and test the re lationships between two characteristics these will be dis cussed in Chapter 7 2 Know when to use which uni or bivariate statistical method Proper use of SPSS not only requires you to have some knowl edge of SPSS but also some expertise in the field of statis
29. nly not not better to some better off definitely better off of extent not better off better off Figure 1 1 Model research project Wealth and happiness 3 Things as would like them to be are certainly are not are to some are are not extent not exactly A ENS cece certainly not to some difficult extremely not difficult difficult extent not difficult difficult 5 I lonely certainly do do not do to some feel definitely not feel eel extent not eel feel Since happiness is not only dependent on financial means but also on other factors the researcher also asks about a number of easily measured characteristics such as the sex age edu cational level and marital family status of the persons being surveyed Sex male female Age in years Marital family statu alleen met partner met partner en kinderen Highest completed educational level Lower secondary education Upper secondary education Higher education The researcher refined his general research question into a num research questions Aun PWN FE ber of specific research questions How many people own cars homes or DVD players How many people are covered by the national health plan How many people receive a rent subsidy How happy are people about their lives How lonely
30. nt role in inductive statistics Even if you compare the means of two sam ples there is some question about the probability that any detect ed difference in means is representative of the population Consider the possibility that the sample of employees from the Labour company consists of 99 women and 99 men You find that the mean IQ of the women is 101 2 and that of the men is 98 2 Can you then state that the female employees of the Labour company are on average more intelligent than the male employ ees You can test if this difference is significant In Chapter 6 we explain how to do that when comparisons are involved and in Chapter 7 how you do that when your research involves a corre H1 How do prepare myself to work with this Introduction significance one or two tailed test one tailed test degrees of freedom df lation When is it then possible to speak of significance The gen eral rule is that a finding is significant if the probability of error p value is less than 5 or with larger samples gt 1000 less than 1 Often SPSS also indicates if a one or two tailed test was involved You conduct a one tailed test when you have formulated a hypothesis or expectation If you have a theory on the basis of which you might expect that female employees are more intelli gent than their male counterparts you can then perform a one tailed test If you do not however have any idea about the possi bi
31. o use a sample in order to estimate a population mean for example The significance of an observed difference or correlation in a sample depends on the number of degrees of freedom which are in turn except in the case of a contingency table often dependent on sample size When you have established that there is for example a signifi cant difference in mean IQ between men and women you would often also like to know the extent to which the difference can be explained predicted To what degree can for example difference in mean IQ be attributed to sex For this purpose there exists de pending on the manner in which the difference is tested all sorts of measures that are often however not included in SPSS We will explain how you can determine the effect size for each test ing method that we discuss 1 3 A few general statistical terms 29 30 data editor 1 4 How does SPSS for Windows work Once SPSS has been properly installed on your computer you should see the following icon on your screen SPs SPSS 12 0 1 for Wind Spss 12 0 1 for Windows Figure 1 3 The SPSS icon You open SPSS by double clicking this icon If you are unable to find the icon you can always click the start button in the bot tom left corner of your screen You then use the mouse indicator to select Programs You can then see if SPSS appears on the program list displayed on screen If such is the case you can start SPSS by double cl
32. own as the alpha reliability index you can see if and if so the extent to which they all in fact measure the same characteristic Using an item analysis you can see if there are questions or items that have a negative effect on the homogeneity If such is the case you can decide not to include responses to such questionnaire elements in the total score for a concept In Chapter 3 How do I modify and combine data we examine among other things how you derive a total score from the scores of the separate items In Chapter 4 we use an example to provide a detailed answer to the question How do I check the homogene ity of the collected scores 1 2 How do I analyze my data A data use key The answers to the following questions are important for the choice of method of statistical analysis 1 Does the general research question involve issues of frequency how often to what extent comparison or correlation Or does it involve a combination of the above 2 What is the level of measurement nominal ordinal or inter val ratio level for the data that you have collected 3 Is a sample or a population involved The block diagram How do I analyze my data to be found on the inside front cover of the book and as a separate insert has been developed on the basis of these questions The following sections explore the issue in greater detail 1 2 1 What specific research questions are involved in my general research question
33. reparation analysis and description of the research data analyzed using SPSS Referring mostly to a study on the relationship between wealth and happiness see Figure 1 1 we will investigate step by step e How you should prepare research data for input Section 1 1 and Chapter 2 How you can use SPSS to adjust and modify your data Before you begin the analysis you often have first to convert recode the values of certain data or combine the values of data to produce new scores Section 1 1 and Chapter 3 and sub sequently check their reliability Section 1 1 and Chapter 4 H1 How do prepare myself to work with this Introduction Will the entire population be involved in my research or will I Prior knowledge No specific prior knowledge is required Before explaining the uses of SPSS we will first introduce the practice example Wealth and happiness see Figure 1 1 e How you choose the appropriate method of analysis To choose the appropriate method of analysis you first have to establish the nature of the research question Section 1 2 1 Does it involve frequencies Chapter 5 comparisons Chapter 6 or correlations Chapter 7 You then must determine the level of measurement for your data Section 1 2 2 Finally you have to decide if a sample or a population is involved in your research You can then use the block diagram How do I analyze my data to ascertain which method of analysis
34. results of a one way analysis of variance one way ANOVA 150 How do analyze my data when a correlation research question is involved 153 Introduction 154 Correlation between two nominal variables Cram r s V 159 What is Cram r s V 159 When do I use Cram r s V 160 How do I calculate Cram r s V 160 How should I read the Cram r s V output 162 How should I report the Cram r s V output 162 Correlation research questions involving ordinal variables Spearman s rank correlation 163 What is Spearman s rank correlation coefficient rho 163 When do I use Spearman s rank correlation 164 How do I calculate a Spearman s rank correlation 165 How should I read the results of a Spearman s rank correlation 167 72 5 7 3 7 3 1 7 3 1 1 7 3 1 2 73 13 1 32 73 201 73 2 2 7 3 2 3 7 3 2 4 L325 1 3 3 7 3 3 1 7 3 3 2 7T 3 3 3 7 3 3 4 7 3 3 5 7 3 4 7 3 4 1 7 3 4 2 7 3 4 3 7 3 4 4 7 3 4 5 7 3 4 6 7 3 4 7 7 3 4 8 7 3 4 9 7 3 4 10 How should I report the results of a Spearman s rank correlation 168 Correlation questions involving interval and ratio variables 168 A scatter plot scattergram and a regression line 168 What is a scatter plot scattergram and a regression line 168 When do I use a scatter plot scattergram and a regression line 169 How do I construct a scatter plot scattergram and a regression line 169 Correlations involving interval and ratio vari
35. rtainly certainly do man 45 00 with partner yes yes no would not are to som are not is to some feel woman 27 00 alone h no no noj yes yes would to so are are nat is to some certainly do woman 30 00 with partner h _ no yes no no no would are not are are not certainly do woman 34 00 with partner I yes yes yes no no would defini are definitel are are not do not feel man 48 00 with partner u no no yes no no would to so are certainl are is to some feel woman 44 00 with partner u yes no no yes yes would are are to som are not certainly do man 23 00 with partner u EE yes yes yes yes no would are are to som are not certainly do man 41 00 with partner _ yes no yesl no sno would arenot arelis certainly do to some woman 47 00 with partner u 10 noj no yes yes yes would certa are aretosom is donotfeel man 2 00 alone T yes yes yes no no would are certain are is certainly certainly do man 44 00 with partner u yes no yes na would are not are are not certainly do woman 40 00 with partner yes yes yes na no would defini are definitel are are not do not feel man 47 00 with partner u 14 yes no no yes yes would are are to som are not certainly do man 28 00 with partner u yes yes yes yes no would are are to som are not certainly do man 42 00 with partner u 6j yes no yes na no would are not are lis certainly do to some woman 46 00 with pa
36. rtner h no no yes yes yes would certa are are to som is do not feel woman 32 00 alone u 18 yes yes yes no no would are certainl are is certainly certainly do man 45 00 with partner h no no yes yes no would not are to som are not is to some feel woman 27 00 alone no no no yes yes would to so are are not is to some certainly do woman 30 00 with partner i yes yes yes no no would defini are definitel are are not do not feel man 48 00 with partner l no no yes no na would to so are certain are is to some feel woman 37 00 with partner h yes no no yes yes would are are to som are not certainly do man 31 00 with partner u yes yes yes yes no would are are to som are not certainly do man 35 00 with partner u yes no yes no no would are not are is certainly do to some woman 35 00 with partner I 26 no nol yes yes would certa are are to som _is do not feel woman 25 00 alone u yes yes yes no no would are certainl are lis certainly certainly do man 37 00 with partner u no no yes yes would not are to som are not is to some feel woman 27 00 alone u no no no yes yes would to sol e are not is to some certainly do woman 30 00 with partner h no yes no no no would are not are are not certainly do woman 39 00 i l del clef dafimital aral nun me j da nat fanl mn AO nn ls SPSS Processor is ready Figure 1 5 The screen that
37. saving your data There are additionally help buttons for providing information about specific procedures involved in the various SPSS tasks themselves kal z ff Verbergen Terug Volgende Soppen Vernieuwen Opties Inhoud Index zoeken ped Geel ven betrroord op _ iO SPSS for Windows provides powerful statistical analysis and data management system beck in a graphical environment using descriptive menus and simple dialog boxes to do most of the work for you Most tasks can bo accomplished simply by pointing and checking the dota files MOUSE database file queries SPSS data files In addition to the simple point and click interface for statistical analysis SPSS for saving output Windows provides drat output format Data Editor A versatile spreadsheet like system for defining entering editing and HTML format displaying data password protecton publishing to Web saving draft output as text tent format Word forest publishing to Web saming pant table views scale findings that can get lost in standard reports Compare groups easily by splitting your in Multidimensional Scali i in Reliability is ng table so that only one group is displayed at a time Beck ain High resolution graphics High resolution full color pie charts bar charts histograms meee seatterplots 3 D graphics and more are meluded as standard features in SPSS dorved mns mimber fomai Database access Retrieve information from databases by using the Database Wizar
38. stion we discuss a few suitable graphic techniques that are available in SPSS The learning process is illustrated on the basis of a sample pro vided for practice entitled Wealth and Happiness or Can money buy happiness Nearly all the problems encountered during the processing and analysis of data are explained by means of this practice example How can this book be used When you have conducted research and obtained a set of numer ical values you can use this book as a guide to the analysis of the data It provides both information about statistics and instruc tions for using SPSS to analyze and interpret the figures In fact it is a book on statistics combined with an SPSS user manual In principle you do not need a separate statistics textbook and need not have any previous knowledge of the subject What do you need In writing this book we have made use of SPSS for Windows Ver sion 12 0 This guide can however also be used with earlier ver sions of SPSS for Windows After you have worked through this book you will be capable of quickly entering simple research data in a computer and ade quately analysing them You will likely also be able to perform statistical techniques that are not presented in this book but with which you are already familiar The menus provide a great deal of information along with relevant examples 10 Where do find the sample data files used in this book Statistical terms and m
39. the factors involved Sound decisions stem from conclusions based on outcomes de rived from research This involves the collection and analysis of data as well as the reporting and dissemination of information The ultimate foundation is provided by the data Figures and sta tistics are objective verifiable and therefore convincing Such views are currently becoming more widespread Just open any given business magazine and you will find articles on this subject all of them concerning the process of taking sound deci sions In this context frequent use is made of such expressions as Data Mining Business Intelligence and Statistical Analy sis What these notions have in common is the fact that they all involve data processing the search for trends problems solu tions or new market opportunities Furthermore they all pro mote the value of considering both subjective and objective fac tors whenever taking decisions With some pride we can say that SPSS is by far the most frequent ly used statistical analysis program The Introduction to Statistics with SPSS constitutes the logical successor to Statistiek met SPSS voor Windows a Dutch publication that appeared in 1999 As you will discover SPSS has taken a further step forward in the areas involving ease of use and professionalism with the development of SPSS 12 0 SPSS can now be used to discover relationships and pat terns to detect causes and to make well
40. tics In the first chapter the basic concepts will be introduced such as level of measurement normal distribution probability signifi cance one and two tailed tests power and effect size Chapters 5 6 and 7 will reveal that the choice of statistical technique is not only dependent on the nature of the research question but also on the level at which the relevant characteristic is measured and the issue of whether a population or a sample is involved Since this book only presents the statistical information that is strictly necessary to analyze data with SPSS using predominantly uni and bivariate analytical methods we will only deal with a few multivariate techniques of analysis 3 Know the conditions and SPSS procedure involved in a given technique Using both explanation and example we detail the essential fac tors in each technique without going too deeply into the underly ing mathematics We additionally indicate when a technique should and should not be used Finally there is an explanation of how to perform the technique using SPSS 4 Be able to read and to interpret SPSS output and to issue re ports based on it The material will be further presented in terms of the following questions which you the user might pose e If the chosen method of statistical analysis has been success fully performed using SPSS how do I read the computer out put What is precisely revealed by the results of the data analy sis e
41. tput in my text 190 Index 193 17 1 How do prepare myself to work with this ntroduction For I don t care too much for money money can t buy me love John Lennon amp Paul McCartney 1 How do I prepare myself to work with this Introduction 1 1 12 z 1 4 How do I prepare the data for How do I analyse my data ew general statistical terms How does SPSS for Windows analysis A data use key work 20 In conducting research the choice of the appropriate statistical technique to analyze the collected data is an important link in the long chain of decisions that must be taken Ultimately the goal of collecting and analysing data is to answer the research question or questions contained in the inquiry To draw attention to the place of data analysis in the research cycle as a whole the typical stages of research are listed below Each stage in the research cycle is stated in the form of a question 1 What is the goal of my research and what is the type of the inquiry 2 How will I acquire information among other ways by reviewing the literature 3 What type of research am I going to conduct 4 What will my research design look like n select a sample Which data collection method am I going to use How will I prepare my data for analysis How will I analyze my data O GO N BW How will I report and evaluate my research This book focuses on stages 7 and 8 along with a portion of 9 p
42. uestion correlation research question 1 2 2 NRN nnn level of measurement nominal level in SPSS Nominal ordinal level in SPSS Ordinal interval level in SPSS Scale ratio level in SPSS Scale continuous variables discrete variables 2 3 population descriptive statistics sample inductive statistics inferential statistics units H1 How do prepare myself to work with this Introduction EEEN nnn w normal distribution Gauss curve standard error sample size homogeneity significance probability probability of error p value alpha a significant statistical testing one or two tailed testing degrees of freedom df effect size A data file data matrix data editor analyze graphs help function tutorials
43. y possible to say that 20 kilos is twice as heavy as 10 kilos In the case of temperature 0 degrees does not after all constitute the natural zero point and consequently an interval level of mea surement is involved Despite this distinction SPSS groups the interval and ratio levels of measurement under the heading of Scale The two other levels of measurement are named Nominal and Ordinal We additionally distinguish between continuous and discrete variables Continuous variables allow you to present a line on which values form a row of connected points a continuum No matter how close to each other any two given points are always separated by still infinitely more other possible values Exam ples of continuous variables are a person s height age and intelli gence Variables that can only have whole values are called dis crete variables such as the number of cars that someone can own or the number of children in a family 1 2 3 Is a population or a sample involved There are two types of statistics descriptive and inferential sta tistics Descriptive statistics are used when research is conduct ed on a population It is possible to speak of a population when all the units about which you wish to make statements are involved in your research That is when you conduct for example a sur vey of all the employees in a company in order to determine their level of job satisfaction H1 How do prepare myself to work with this
44. yping your data into the data editor which appears on the screen when you start SPSS It is additionally helpful if you immediately assign a name to the variables for which you are entering the scores Before you begin the analysis of your data you must check if values or codes are included that do not belong there For sex use is most frequently made of the codes 1 male and 2 female The entries under sex in your database must therefore only be ones and twos If for example a seven appears you have probably made a mistake when entering the data You need to correct it before continuing with your analysis Similarly you have to check if the variables display sufficient variation If there are very few men in your sample it will be dif ficult for you to determine if a difference exists between men and women in for example the degree of their happiness We investi gate this further in Section 2 4 How do I check if I have made a mistake when entering my data and again in Chapter 5 when we deal with the calculation of frequencies Sometimes you have to process the data that you have entered We are for example not only interested in a person s desire to live life over again or the fact that someone may find life diffi cult but above all the extent to which an individual agrees with such expressions of un happiness By adding up the scores from the separate items in the questionnaire we can arrive at
Download Pdf Manuals
Related Search