INTRODUCTION TO QUANTITATIVE RESEARCH METHODS CHAPTER 1 ppsx

# Mark Balnaves and Peter Caputi 2001Introduction # Alec McHoul 2001 First published 2001 Apart from any fair dealing for the purposes of research or private study, or criticism or revie

I N T R O D U C T I O N T O

Q U A N T I T A T I V E

R E S E A R C H M E T H O D S

I NTRODUCTION

TO QUANTITATIVE

RESEARCH METHODS

# Mark Balnaves and Peter Caputi 2001

Introduction # Alec McHoul 2001

First published 2001

Apart from any fair dealing for the purposes of research or private study, or criticism or review, as permitted under the Copyright, Designs and Patents Act, 1988, this publication may be reproduced, stored or transmitted in any form, or by any means, only with the prior permission

in writing of the publishers, or in the case of reprographic reproduction, in accordance with the terms of licences issued by the Copyright Licensing Agency Inquiries concerning reproduction outside those terms should be sent to the publishers.

SAGE Publications Ltd

6 Bonhill Street

London EC2A 4PU

SAGE Publications Inc

2455 Teller Road

Thousand Oaks, California 91320

SAGE Publications India Pvt Ltd

32, M-Block Market

Greater Kailash-I

NewDelhi 110 048

British Library Cataloguing in Publication data

A catalogue record for this book is

available from the British Library

ISBN 0-7619-6803-2

ISBN 0-7619-6804-0 (pbk)

Library of Congress catalog record available

Typeset by Keyword Publishing Services Limited, UK Printed in Great Britain by The Cromwell Press Ltd, Trowbridge, Wiltshire

2 Starting the Inquiry: `But what happened then?' 10

Exploration, Description and Explanation 16

Collecting Data Across Cultures: Can we measure cultural variation?Culture's Consequences (Geert Hofstede) 51

Hypotheses and Operationalization 53

v

External Validity and Sampling 90Population and Sampling Frame 91

Stratified Random Sampling 92Multi-Stage Cluster Sampling 92

Non-Probability Sampling 95

Great Media and Politics Detective Stories:

Using Survey Data Do media change people's political attitudes?

The People's Choice (Paul Lazarsfeld) 96

Hypotheses and Operationalization 98

vi

vii

Using SPSS: Correlation and Regression 163

Looking at CategoricalData 171Exploring Bivariate Categorical Data 171Example 6.3: Computing the contingency coefficient 174Inference: From Samples to Populations 175Parameters,Estimates and Statistics 176Sampling Distributions 177Hypothesis Testing ± Don't just show me the evidence,convince me

Is There An Ideal Person?: Quetelet,Galton,Pearson 214Does a General Intelligence Factor Exist?:

Spearman and the introduction of correlation 215

Is There a Genetic Component to Intelligence?

Did Burt clone the data? ± data ain't always data!! 217Are Social Factors More Important than Individual Inclinations?

Durkheim's study on suicide (The use of secondary data) 218

Appendix I: Sample Letter for Informed Consent 237Appendix II: BSA Statement of EthicalPractice 239Appendix III: The StatisticalInquirer 246

C O N T E N T S

viii

Table 2.1 Impact of television on Western Samoan evening activities,

Table 2.2 Time and research design 26

Table 2.3 Organizing a quantitative research study 28

Table 3.1 From construct to operational definition 54

Table 3.2 Actual questions used to construct individualism/masculinity

Table 3.3 Countries ranked by individualism scores 56

Table 3.4 Countries ranked by masculinity scores 56

Table 3.5 Countries ranked by power distance scores 57

Table 3.6 Countries ranked by uncertainty avoidance scores 58Table 4.1 Effectiveness of Program A and B for men with initially

unfavourable and men with initially favourable attitudes 74Table 4.2 Effectiveness of Program A and B for men of different

educational backgrounds 74

Table 4.3 Examples from Thurstone's differential scale 80

Table 4.4 Examples from Bogardus Social Distance scale 80

Table 4.5 Selected examples from Christie and Geis's Likert scale 81Table 4.6 Sample questionnaire and coding column 86

Table 4.7 Rank order of French high school students' intelligence scores

obtained with three hypothetical measures 89

Table 4.8 Table of random numbers 92

Table 4.9 Random selection of households for interview 94

Table 4.10 Summary of sampling procedures 94

Table 5.1 Example of back-to-back plot 113

Table 5.2 Frequency distribution table for grouped data 114

Table 5.3 Hypothetical data for variables X and Y 117

Table 5.4 The anatomy of a table 120

Table 5.5 Mean ratings of intensity of emotion 122

Table 5.6 Reframed data: mean ratings of intensity of emotion 122Table 5.7 Accounting for tastes: comparison of stratified sample with

official statistics 124

Table 6.1 Hypothetical data 149

Table 6.2 Hypothetical data on correlation between years of counselling

experience and effective outcome 154

Table 6.3 Data with outliers 156

Table 6.4 Distribution of votes by section 172

Table 6.5 Percentage of votes within each section 172

Table 6.6 A 2  2 contingency table 173

Table 6.7 Data for job performance and EC test 174

Table 6.8 Computational details for Table 6.7 175

Table 6.9 Some commonly used values from a set of normal tables 182Table 6.10 Frequencies in colour and shape of peas 184

Table 6.11 Hypothetical data for number of successful free throws in two

conditions 185

Table 6.12 Descriptive statistics 187

Table 6.13 Frequencies relating to marital status and IQ 192

Table 6.14 Comparison of observed and expected frequencies for

preferences by gender 233

TA B L E S

x

Chart 1.1 Reports on Aboriginal youth and all youth-crime reports,

Feb 1991 to Jan 1992 3

Chart 1.2 News reports on Aboriginal youth-crime and all youth-crime

with actual crime data, Feb 1991 to Jan 1992 4

Figure 2.1 From assertions to evidence 23

Figure 4.5 Different answering formats 84

Figure 4.6 Contingency questions 84

Figure 4.7 Factors affecting people's motivation to provide complete and

accurate information to the interviewer 87

Figure 4.8 Checklist for questionnaire design 88

Figure 4.9 Multi-stage cluster sampling ± following the census

tracts 93

Figure 4.10 Experimental choice based on issues of internal and external

validity 96

Figure 4.11 Whereas actual occupation does little to refine the relationship

between SES level and vote, it makes more difference whether

a voter considers himself as belonging to `business' or

`labour' 100

Figure 4.12 Religious affiliation splits the vote sharply 101

Figure 4.13 One-step model of mass-media influence 102

Figure 4.14 Two-step model of mass-media influence 102

Figure 4.15 Inductive approach 104

Figure 4.16 Deductive approach 104

Figure 5.1 A stem and leaf display of ESP data 114

Figure 5.2 Histogram of hypothetical examination marks 115Figure 5.3 The anatomy of a boxplot 116

Figure 5.4 Boxplots for two hypothetical variables X and Y 117Figure 5.5 Side-by-side boxplots 118

Figure 5.6 Preference for telecommunications carrier 121

Figure 5.7 A different way to display preference for telecommunications

carrier 121

Figure 6.1 Scatterplot for data in Table 6.1 150

Figure 6.2 Negative association between two variables 150Figure 6.3 No relationship between two variables 151

Figure 6.4 Scatterplot showing an outlier 151

Figure 6.5 A curvilinear relationship 155

Figure 6.6 Fitting a line through points on a scatterplot 157Figure 6.7 Illustrating the concept of residual 158

Figure 6.8 Bar chart of frequency of votes by selection 172Figure 6.9 Bar chart of raw percentage of votes by selection 173Figure 6.10 The normal distribution of a normal curve 179Figure 6.11 The t-distribution 183

Figure 6.12 Durkheim's theoretical hypothesis on suicide 220Figure 6.13 Durkheim's hypothesis of degree of integration 221Figure 6.14 Durkheim's hypothesis of degree of regulation 222

F I G U R E S

xii

A special thanks to:

Gary Bouma and David DeVaus

Mark Busani, Nick Castle and Monica Vecchiotti for their help with themultimedia courseware

Maurice Dunlevy, for contributions on journalism, and Harry Oxley, forcontributions on causal diagramming

Patrick Rawstorne for use of his PhD dataset Predicting and Explaining theuse of Information Technology with Value Expectancy Models of Behaviour

in Contexts of Mandatory Use

Erika Pearson ± it's hard to find good help nowadays

Tony Bennett, Mike Emmison and John Frow for use of their dataset fromThe Australian Everyday Consumption project

SPSS illustrations have been reprinted by SPSS copyright permission Excelillustrations have been reprinted by Excel copyright permission

The Apple University Consortium and PCTech for their equipment andsoftware support

The Australian National Library for assistance with access to The Strand,from which the original Paget sketches of Sherlock Holmes were repro-duced for this book

Alec McHoul, Mike Innes, Joyce and Michele Balnaves, Wendy Parkins andJames Donald, who provided valuable insights into detection

John and Paul Balnaves on questions of Shakespeare and logic

Michele, Mary-Claire, Gerard, Elayne, James, and Jack

Douglas Adams ± the bottle of red has been sent

The authors of detective fiction

The authors and publishers wish to thank the following for their permission

to use copyright material:

Tables 6.16 and 6.17 Copyright # 1993 Canadian PsychologicalAssociation Reprinted with permission

Page 12, extract from The Strange Crime of John Boulnois by G.K Chesterton.Used by permission of A.P Watt on behalf of The Royal Literary Fund.Page 14, extract from Dirk Gently's Holistic Detective Agency by DouglasAdams Used by permission of Douglas Adams Copyright # 1987.Heinemann

Figure 2.1 From Evaluating Social Science Research, second edition, by Paul C.Stern and Linda Kalof Copyright # 1979, 1996 Oxford University Press,Inc Used by permission of Oxford University Press

Table 2.2 Used by permission of Gary Bouma Copyright # 1993 TheResearch Process Oxford University Press

Page 38, extract from The Blue Cross by G.K Chesterton Used by permission

of A.P Watt on behalf of The Royal Literary Fund

Figure 4.3 Used by permission of David DeVaus Copyright # 1990.Surveys in Social Research Allen and Unwin

Table 4.7 From Research Methods in Social Relations, sixth edition, by Charles

M Judd, Eliot R Smith and Louise H Kidder Copyright # 1991, Holt,Rinehart and Winston Reproduced by permission of the publishers.Every effort has been made to trace all copyright holders, but if any havebeen overlooked, or if any additional information can be given, the publish-ers will be pleased to make the necessary amendments at the first oppor-tunity

Order at All Points

Counting and accounting

A man is driving through the bush one day and has to stop while a farmer takes his sheep across the road There are quite a lot of sheep, so it takes a fair while When they've all passed by, the man goes up to the farmer and asks, `If I can tell you how many sheep you have, to within one either way, can I have one of them?' The farmer replies, `Course you can You'll never get it right.' The man says, `You have six thousand four hundred and twenty two.' `Well blow me down,' replies the farmer ± or words to that effect `In fact I have six thousand four hundred and twenty one I counted them this morning.' So the man walks back to the car with his prize.

`Wait on,' cries the farmer `If I can tell you what your job is, can I have her back?' `Sure,' says the man, `You'll never guess.' `Well,' says the farmer, `I figure you'd be a statistician with the Australian Bureau of Statistics.' `Well I'll be !' the man replies, `Exactly right How on earth did you know that?'

The farmer comes back: `Put me dog down and I'll tell you.'

Traditional Australian Bush Yarn

THE Ql-Qt CONTINUUM1

Like many in the humanities and social sciences, I was trained to be (at theleast) sceptical about statistical methods and (at most) downright hostiletowards them In sceptical mode, I was exhorted to use statistics not in theway a drunk uses a lamppost: for support rather than illumination Inhostile mode, the word was that statistics was for `positivists' (a very unfaircharacterization, as it turns out, of positivism) What all of this well-mean-ing and humanistic advice ignored was the sheer fact that our social andcultural worlds, today, are massively subject to statistical accounts (seeHacking, 1982) Whenever we turn on the TV news or open a newspaper,the world is now routinely accounted for in terms of the numbers it gen-erates: from world population statistics right down to chewing gum mar-kets In this respect, it's not quite as if numbers were on one side of the coinand `lived cultures' on the other Rather, the technologies of numberinghave become just one (though, in some disciplines, a dominant one) ofthe many practices that make up the cultures of modernity In this briefintroduction, then, I want to think through the supposed distinction (binary,even) between the quantitative (Qt) and the qualitative (Ql) and to showthat the seal between the two is by no means as watertight as it is oftenassumed to be

My first realization of an elision between Qt and Ql came to me when, out

of sheer impecuniousness, I went to work for the Survey Research Centre atthe Australian National University (ANU) in the mid-1970s Prior to thisway of supplementing my meagre PhD scholarship, my only encounterwith statistics had been the compulsory undergraduate methods course

in sociology, taught, as it happened, by a died-in-the-wool symbolic actionist, a Ql-man if ever there was one! Said lecturer was, then, veryhappy for me to complete my statistics assignments by having a friendwho was a physics student crunch the exercises on the university's onemainframe computer by submitting bundles of punchcards Not, then,exactly the best of trainings or qualifications, I admit But working late atANU, designing and administering the Australian Capital Territory (ACT)population surveys, I came to see what a symbolic and interactional process

inter-Qt work could be in practice One of our clients at the time was the localFamily Planning organization It wanted to know which forms of contra-ception were most in use in the Capital Territory The only problem withthis was that the official sampling procedures required interviewers tojointly interview two members of each household selected (using lot num-bers) on a rotational basis: oldest and third oldest in odd-numbered lots,and second oldest and fourth oldest in even-numbered lots This meant, ineffect, that a fair proportion of interviews involved parents and their olderchildren ± not exactly the best interactional setting to ask people about theircontraceptive practices The problem was both, and equally, statistical and

`cultural.' Qt and Ql could not be a simple binary And, oh yes, the ful `solution' we developed was to draw up a card with each kind of contra-ception numbered Respondents were then shown the card and would saysuch things as `Well, I tried the number seven but it didn't work for me, sonow I prefer the twenty six.'

wonder-The same realization came back to me during a more recent researchproject (Mickler and McHoul, 1998) In this project, we collected over 600newspaper articles on Aborigines, youth and crime over a 12-month period

in the early 1990s in order to see whether there had been, as some suspected

at the time, a media-generated `crime wave.' We had a neutral reader/research-assistant type the articles into a relational database program(QSR NUDEIST) and, at the same time, code the articles for such things

as `source' (the origin of the reported events), `participants' (the categories

of persons reported on in each article) and how the reader thought thearticle was treating such `participants' (in positive, negative or neutralmoral terms) What we hoped to get out of this was a strongly Ql argumentbased on a discursive analysis of the news articles and their `readings.'However, before long, we found that working with over 600 texts wouldnot allow us to do this The data in question were simply too numerous.And anyway, NUDEIST was starting to generate matrices of such things as

`Date of publication'  `Article source' and `Newspaper'  `Participants.'Each cell of the matrix listed the relevant articles by their uniqueNUDEIST document number There was no way we could work with this

B A L N AV E S A N D C A P U T I

2