First published 2010 - Amazon Web Servicestandfbis.s3.amazonaws.com/rt-media/pp/common/... · Part...

First published 2010by Psychology Press27 Church Road, Hove, East Sussex BN3 2FA

Simultaneously published in the USA and Canadaby Psychology Press270 Madison Avenue, New York, NY 10016

Psychology Press is an imprint of the Taylor & Francis Group,an Informa business

Copyright © 2010 Psychology Press

Typeset in Palatino by RefineCatch Limited, Bungay, SuffolkPrinted and bound in Great Britain byTJ International Ltd, Padstow, CornwallCover design by Lisa Dynan

All rights reserved. No part of this book may be reprinted orreproduced or utilised in any form or by any electronic, mechanical orother means, now known or hereafter invented, including photocopyingand recording, or in any information storage or retrieval system, withoutpermission in writing from the publishers.

British Library Cataloguing in Publication DataA catalogue record for this book is available from the British Library

Library of Congress Cataloging-in-Publication DataClark-Carter, David.

Quantitative psychological research : a student’s handbook /David Clark-Carter. – 3rd ed.

p. cm.Includes bibliographical references and index.1. Psychology—Research—Methodology—Textbooks. I. Title.

BF76.5.C53 2009150.72—dc22

2009006100

ISBN: 978–1–84169–690–4 (hbk)ISBN: 978–1–84169–691–1 (pbk)

Copyright Psychology Press 2010 - http://www.psypress.com/9781841696911

http://www.psypress.com/9781841696911

Contents Detailed contents of chapters ix

Preface xiv

Part 1Introduction 1

1 The methods used in psychological research 3

Part 2Choice of topic, measures and research design 19

2 The preliminary stages of research 213 Variables and the validity of research designs 374 Research designs and their internal validity 49

Part 3Methods 69

5 Asking questions I: Interviews and surveys 716 Asking questions II: Measuring attitudes and meaning 867 Observation and content analysis 98

Part 4Data and analysis 107

8 Scales of measurement 1099 Summarising and describing data 116

10 Going beyond description 14211 Samples and populations 15112 Analysis of differences between a single sample and a population 16113 Effect size and power 17914 Parametric and non-parametric tests 18715 Analysis of differences between two levels of an independent variable 19716 Preliminary analysis of designs with one independent variable with more than

two levels 22117 Analysis of designs with more than one independent variable 24318 Subsequent analysis after ANOVA or χ2 25919 Analysis of relationships I: Correlation 284

viiCopyright Psychology Press 2010 - http://www.psypress.com/9781841696911


20 Analysis of relationships II: Regression 31421 Analysis of covariance (ANCOVA) 33922 Screening data 35723 Multivariate analysis 36424 Meta-analysis 377

Part 5Sharing the results 389

25 Reporting research 391

AppendixesI. Descriptive statistics 413

(linked to Chapter 9)II. Sampling and confidence intervals for proportions 423

(linked to Chapter 11)III. Comparing a sample with a population 428

(linked to Chapter 12)IV. The power of a one-group z-test 434

(linked to Chapter 13)V. Data transformation and goodness-of-fit tests 437

(linked to Chapter 14)VI. Seeking differences between two levels of an independent variable 444

(linked to Chapter 15)VII. Seeking differences between more than two levels of an independent variable 468

(linked to Chapter 16)VIII. Analysis of designs with more than one independent variable 490

(linked to Chapter 17)IX. Subsequent analysis after ANOVA or χ2 505

(linked to Chapter 18)X. Correlation and reliability 522

(linked to Chapter 19)XI. Regression 541

(linked to Chapter 20)XII. ANCOVA 558

(linked to Chapter 21)XIII. Evaluation of measures: Item and discriminative analysis, and accuracy of tests 560

(linked to Chapter 6)XIV. Meta-analysis 564

(linked to Chapter 24)XV. Probability tables 577XVI. Power tables 617XVII. Miscellaneous tables 661

References 671

Glossary of symbols 677

Author index 678

Subject index 680

viii Contents



THE METHODS USEDIN PSYCHOLOGICALRESEARCH 1IntroductionThis chapter deals with the purposes of psychological research. It explainswhy psychologists employ a method in their research and describes the rangeof quantitative methods employed by psychologists. It addresses the ques-tion of whether psychology is a science. Finally it deals with ethical issues todo with psychological research.

What is the purpose of research?The purpose of psychological research is to increase our knowledge ofhumans. Research is generally seen as having one of four aims, which canalso be seen as stages: the first is to describe, the second is to understand,leading to the third, which is to predict, and then finally to control. In the caseof research in psychology the final stage is better seen as trying to intervene toimprove human life. As an example, take the case of non-verbal communica-tion (NVC). Firstly, psychologists might describe the various forms of NVC,such as eye contact, body posture and gesture. Next they will try to under-stand the functions of the different forms and then predict what will happenwhen people display abnormal forms of NVC, such as making too little eyecontact or standing too close to others. Finally they might devise a means oftraining such people in ways of improving their NVC. This last stage will alsoinclude some evaluation of the success of the training.

What is a method?A method is a systematic approach to a piece of research. Psychologists use awide range of methods. There are a number of ways in which the methodsadopted by psychologists are classified. One common distinction which ismade is between quantitative and qualitative methods. As their names sug-gest, quantitative methods involve some form of numerical measurementwhile qualitative methods involve verbal description.



Why have a method?The simple answer to this question is that without a method the research of apsychologist is no better than the speculations of a layperson. For, without amethod, there is little protection against our hunches overly guiding whatinformation is available to us and how we interpret it. In addition, withoutmethod our research is not open to the scrutiny of other psychologists. As anexample of the dangers of not employing a method, I will explore the ideathat the consumption of coffee in the evening causes people to have a poornight’s sleep.

I have plenty of evidence to support this idea. Firstly, I have my ownexperience of the link between coffee consumption and poor sleep. Secondly,when I have discussed it with others they confirm that they have the sameexperience. Thirdly, I know that caffeine is a stimulant and so it seems aperfectly reasonable assumption that it will keep me awake.

There are a number of flaws in my argument. In the first place I know myprediction. Therefore the effect may actually be a consequence of that know-ledge. To control for this possibility I should study people who are unawareof the prediction. Alternatively, I should give some people who are aware ofthe prediction what is called a placebo—a substance which will be indis-tinguishable from the substance being tested but which does not have thesame physical effect—in this case a drink which they think contains caffeine.Secondly, because of my prediction I normally tend to avoid drinking coffeein the evening; I only drink it on special occasions and it may be that otheraspects of these occasions are contributing to my poor sleep.

The occasions when I do drink coffee in the evenings are when I havegone out for a meal at a restaurant or at a friend’s house or when friendscome to my house. It is likely that I will eat differently on these occasions: Iwill have a larger meal or a richer meal and I will eat later than usual. Inaddition, I may drink alcohol on these occasions and the occasions may bemore stimulating in that we will talk about more interesting things than usualand I may disrupt my sleeping pattern by staying up later than usual. Finally,I have not checked on the nature of my sleep when I do not drink coffee; Ihave no baseline for comparison.

Thus, there are a number of factors which may contribute to my poorsleep, which I need to control for if am going to study the relationshipbetween coffee consumption and poor sleep properly. Applying a method tomy research allows me to test my ideas more systematically and morecompletely.

Tensions between control andecological validityThroughout science there is a tension between two approaches. One is toinvestigate a phenomenon in isolation, or, at least, with a minimum of otherfactors, which could affect it, being present. For example, I may isolate theconsumption of caffeine as the factor which contributes to poor sleep. Thealternative approach is to investigate the phenomenon in its natural setting.

4 Introduction



For example, I may investigate the effect of coffee consumption on my sleepin its usual context. There are good reasons for adopting each of theseapproaches.

By minimising the number of factors present, researchers can exercisecontrol over the situation. Thus, by varying one aspect at a time and observ-ing any changes, they can try to identify relationships between factors. Thus,I may be able to show that caffeine alone is not the cause of my poor sleep. Inorder to minimise the variation which is experienced by the different peoplethey are studying, psychologists often conduct research in a laboratory.

However, often when a phenomenon is taken out of its natural setting itchanges. It may have been the result of a large number of factors workingtogether or it may be that, by conducting my research in a laboratory, I havemade it so artificial that it bears no relation to the real world. The termecological validity is used to refer to research which does relate to real-worldevents. Thus, the researcher has to adopt an approach which maximisescontrol while at the same time being aware of the problem of artificiality.

Distinctions between quantitative andqualitative methodsThe distinction between quantitative and qualitative methods can be a falseone, in that they may be two approaches to studying the same phenomena.Or they may be two stages in the same piece of research, with a qualitativeapproach yielding ideas which can then be investigated via a quantitativeapproach. The problem arises when they provide different answers. Nonethe-less, the distinction can be a convenient fiction for classifying methods.

Quantitative methods

One way to classify quantitative methods is under the headings of experi-menting, asking questions and observing. The main distinction between thethree is that in the experimental method researchers manipulate certainaspects of the situation and measure the presumed effects of those manipula-tions. Questioning and observational methods generally involve measure-ment in the absence of manipulation. Questioning involves asking peopleabout details such as their behaviour and their beliefs and attitudes. Obser-vational methods, not surprisingly, involve watching people’s behaviour.

Thus, in an experiment to investigate the relationship between coffeedrinking and sleep patterns I might give one group of people no coffee,another group one cup of normal coffee and a third group decaffeinatedcoffee and then measure how much sleep members of each group had. Alter-natively, I might question a group of people about their patterns of sleep andabout their coffee consumption, while in an observational study I might staywith a group of people for a week, note each person’s coffee consumptionand then, using a closed circuit television system, watch how well they sleepeach night.

The distinction between the three methods is, once again, artificial, forthe measures used in an experiment could involve asking questions or

1. The methods used in psychological research 5



making observations. Before I deal with the three methods referred to above Iwant to mention a method which is often left out of consideration and givesthe most control to the researcher—modelling.

Modelling and artificial intelligence

ModellingModelling refers to the development of theory through the construction ofmodels to account for the results of research and to explore more fully theconsequences of the theory. The consequences can then be subjected toempirical research to test how well the model represents reality. Models cantake many forms. They have often been based on metaphors borrowed fromother disciplines. For example, the information-processing model of humancognition can be seen to be based on the computer. As Gregg (1986) pointsout, Plato viewed human memory as being like a wax tablet, with forgettingbeing due to the trace being worn away or effaced; see also Randall (2007) fora discussion of metaphors of memory.

Modelling can be in the form of the equivalent of flow diagrams as perAtkinson and Shiffrin’s (1971) model of human memory, where memory isseen as being in three parts: immediate, short-term and long-term. Alter-natively, it can be in the form of mathematical formulae, as were Hull’smodels of animal and human learning (see Estes, 1993). Friston (2005) dis-cusses models of how the brain functions, including statistical models.

With the advent of the computer, models can now be explored throughcomputer programs. For example, Newell and Simon (1972) explored humanreasoning through the use of computers. This approach to modelling is calledcomputer simulation. Miller (1985) has a good account of the nature of com-puter simulation, while Brattico (2008) and Fodor (2000) discuss the limita-tions of current approaches.

Artificial intelligenceA distinction needs to be made between computer simulation and artificialintelligence. The goal of computer simulation is to mimic human behaviouron a computer in as close a way as possible to the way humans perform thatbehaviour. The goal of artificial intelligence is to use computers to performtasks in the most efficient way that they can and not necessarily in the waythat humans perform the tasks. Nonetheless, the results of computer simula-tion and of artificial intelligence can feed back into each other, so that theresults of one may suggest ways to improve the other. See Boden (1987) for anaccount of artificial intelligence.

The experiment

Experiments can take many forms, as you will see when you read Chapter 4on designs of research. For the moment I simply want to re-emphasise thatthe experimenter manipulates an aspect of the situation and measures whatare presumed to be the consequences of those manipulations. I use the termpresumed because an important issue in research is attempting to identifycausal relationships between phenomena. As explained earlier, I may havepoorer sleep when I drink coffee but it might not be the cause of my poor

6 Introduction



sleep; rather, it might take place when other aspects of the situation, whichdo impair my sleep, are also present. It is felt that the properly designedexperiment is the best way to identify causal relationships.

By a properly designed experiment I mean one in which all those aspectsof the situation which may be relevant are being controlled for in some way.Chapter 4 discusses the various means of control which can be exercised byresearchers.

The quasi-experiment

The quasi-experiment can be seen as a less rigorous version of the experi-ment. For example, where the researcher does not manipulate an aspect of thesituation, such as coffee consumption, but treats people as being in differentgroups on the basis of their existing consumption, or lack of it, and thencompares the sleep patterns of the groups. Because the quasi-experiment isless well controlled than an experiment, identifying causal relationships canbe more problematic. Nonetheless, this method can be used for at least twogood reasons: firstly, when it is not possible to manipulate the situation;secondly, it can have better ecological validity than the experimentalequivalent.

Asking questions

There are at least three formats for asking questions and at least three ways inwhich questions can be presented and responded to. The formats areunstructured (or free) interviews, semi-structured interviews and structuredquestionnaires. The presentation modes are face-to-face, by telephone orthrough written questionnaire. Surveys of people usually employ somemethod for asking questions.

Unstructured interviewsAn unstructured interview is likely to involve a particular topic or topics tobe discussed but the interviewer has no fixed wording in mind and is happyto let the conversation deviate from the original topic if potentially interestingmaterial is touched upon. Such a technique could be used when a researcheris initially exploring an area with a view to designing a more structuredformat for subsequent use. In addition, this technique can be used to producethe data for a content analysis (see below) or even for a qualitative methodsuch as discourse analysis (see Potter & Wetherall, 1995).

Semi-structured interviewsSemi-structured interviews are used when the researcher has a clearer ideaabout the questions which are to be asked but is not necessarily concernedabout the exact wording, or the order in which they are to be asked. It is likelythat the interviewer will have a list of questions to be asked in the course ofthe interview. The interviewer will allow the conversation to flow compara-tively freely but will tend to steer it in such a way that he or she can introducespecific questions when the opportunity arises. An example of thesemi-structured interview is the typical job interview.




The structured questionnaireThe structured questionnaire will be used when researchers have a clear ideaabout the range of possible answers they wish to elicit. It will involve precisewording of questions, which are asked in a fixed order and each one of whichis likely to require respondents to answer one of a number of alternativeswhich are presented to them. For example:

There are a number of advantages of this approach to asking questions.Firstly, respondents could fill in the questionnaire themselves, which meansthat it could save the researcher’s time both in interviewing and in travellingto where the respondent lives. Secondly, a standard format can minimise theeffect of the way in which a question is asked on the respondent and on his orher response. Without this check any differences which are found betweenpeople’s responses could be due to the way the question was asked ratherthan any inherent differences between the respondents. A third advantage ofthis technique is that the responses are more immediately quantifiable. In theabove example, respondents can be said to have scored 1 if they said that theystrongly agreed with the statement and 5 if they strongly disagreed.

Structured questionnaires are mainly used in health and social psych-ology, by market researchers and by those conducting opinion polls. Focusgroups can be used to assess the opinions and attitudes of a group of people.They allow discussion to take place during or prior to the completion of aquestionnaire and the discussion itself can be recorded. They can be particu-larly useful in the early stages of a piece of research when the researchers aretrying to get a feel for a new area. Interviews and surveys are discussedfurther in Chapters 5 and 6.

Observational methods

There is often an assumption that observation is not really a method as aresearcher can simply watch a person or group of people and note downwhat happened. However, if an observation did start with this approach itwould soon be evident to the observer that, unless there was little behaviourtaking place, it was difficult to note everything down.

There are at least three possible ways to cope with this problem. The firstis to rely on memory and write up what was observed subsequently. Thisapproach has the obvious problem of the selectivity and poor retention ofmemory. A second approach is to use some permanent recording device,such as audio or video, which would allow repeated listening or viewing. Ifthis is not possible, the third possibility is to decide beforehand what aspectsof the situation to concentrate on. This can be helped by devising a codingsystem for behaviour and preparing a checklist beforehand.

You may argue that this would prejudge what you were going to observe.However, you must realise that even when you do not prepare for an obser-vation, whatever is noted down is at the expense of other things which were

8 Introduction



not noted. You are being selective and that selectivity is guided by someimplicit notion, on your part, as to what is relevant. As a preliminary stageyou can observe without a checklist and then devise your checklist as a resultof that initial observation but you cannot escape from the selective process,even during the initial stage, unless you are using a means of permanentlyrecording the proceedings. Remember, however, that even a video camerawill be pointed in a particular direction and so may miss things.

Methods involving asking questions and observational methods span thequalitative–quantitative divide.

Structured observationStructured observation involves a set of classifications for behaviour and theuse of a checklist to record the behaviour. An early version, which is still usedfor observing small groups, is the interaction process analysis (IPA) devisedby Bales (1950) (see Hewstone & Stroebe, 2001).

Using this technique, verbal behaviour can be classified according tocertain categories, such as ‘Gives suggestion and direction, implying auton-omy for others’. Observers have a checklist on which they record the natureof the behaviour and to whom it was addressed. The recording is done sim-ply by making a mark in the appropriate box on the checklist every time anutterance is made. The IPA loses a lot of the original information but that isbecause it has developed out of a particular theory about group behaviour. Inthis case, the theory is that groups develop leaders, that leaders can be of twotypes, that these two can co-exist in the same group and that interactions withthe leaders will be of a particular type. A more complicated system couldinvolve symbols for particular types of behaviour, including non-verbalbehaviour.

Structured observation does not only have to be used when present at theoriginal event. It is also often used to summarise the information on a videoor audio recording. It has the advantage that it prepares the information forquantitative statistical analysis.

A critical point about structured observation, as with any measure whichinvolves a subjective judgement, is that the observer, and preferably obser-vers, should be clear about the classificatory system before implementing it.In Chapter 2, I return to this theme under the heading of the reliability ofmeasures. For the moment, it is important to stress that an observer shouldclassify the same piece of behaviour in the same way from one occasion toanother. Otherwise, any attempt to quantify the behaviour is subject to error,which in turn will affect the results of the research. Observers shouldundergo a training phase until they can classify behaviour with a high degreeof accuracy. It is preferable to have more than one observer because if theydisagree over a classification this will show that the classification is unclearand needs to be refined further. Structured observation is dealt with inChapter 7.

Content analysis

Content analysis is a technique used to quantify aspects of written or spokentext or of some form of visual representation. The role of the analyst is todecide on the unit of measurement and then apply that measure to the text or




other form of representation. For example, Pitts and Jackson (1989) looked atthe presence of articles on the subject of AIDS in Zimbabwean newspapers, tosee whether there was a change with a government campaign designed toraise awareness and whether any change was sustained. In a separate study,Manstead and McCulloch (1981) looked at the ways in which males andfemales were represented in television adverts. Content analysis is dealt within Chapter 7.

Meta-analysis

Meta-analysis is a means of reviewing quantitatively the results of theresearch in a given area from a number of researchers. It allows the reviewerto capitalise on the fact that while individual researchers may have usedsmall samples in their research, an overview is based on a number of suchsmall samples. Thus, if different pieces of research come to different conclu-sions, the overview will show the direction in which the general trend ofrelevant research points. Techniques have been devised which allow thereviewer to overcome the fact that individual pieces of research may haveused different statistical procedures in producing the summary. A fullerdiscussion can be found in Chapter 24.

Case studies

Case studies are in-depth analyses of one individual or, possibly, one institu-tion/organisation at a time. They are not strictly a distinct method butemploy other methods to investigate the individual. Thus, a case study mayinvolve both interviews and experiments. They are generally used when anindividual is unusual: for example, when an individual has a particular skillsuch as a phenomenal memory (see Luria, 1975a). Alternatively, they areused when an individual has a particular deficit such as a form of aphasia—an impairment of memory (see Luria, 1975b). Cognitive neuropsychologistsfrequently use case studies with impaired people to help understand hownormal cognition might work (see Humphreys & Riddoch, 1987).

Qualitative methods

Two misunderstandings which exist about the qualitative approach toresearch are, firstly, that it does not involve method and, secondly, that it iseasier than quantitative research. While this may be true of bad research,good qualitative research will be just as rigorous as good quantitativeresearch. Many forms of qualitative research start from the point of view thatmeasuring people’s behaviour and their views fails to get at the essence ofwhat it is to be human. To reduce aspects of human psychology to numbersis, according to this view, to adopt a reductionist and positivist approach tounderstanding people.

Reductionism refers to reducing the object of study to a simpler form.Critics of reductionism would argue, for example, that you cannot under-stand human memory by giving participants lists of unrelated words, meas-uring recall and looking at an average performance. Rather, you have to

10 Introduction



understand the nature of memories for individuals in the wider context oftheir experience, including their interaction with other people. Positivismrefers to a mechanistic view of humans which seeks understanding in termsof cause and effect relationships rather than the meanings of individuals. Thepoint is made that the same piece of behaviour can mean different things todifferent people and even to the same person in different contexts. Thus, ahandshake can be a greeting, a farewell, the conclusion of a contest or thesealing of a bargain. To understand the significance of a given piece ofbehaviour, the researcher needs to be aware of the meaning which it has forthe participants. The most extreme form of positivism which has beenapplied in psychology is the approach adopted by behaviourism.

In the first edition of this book I briefly described some qualitativemethods. In subsequent editions I have had a dilemma in that I want toexpand that section to cover some more methods while at the same time Ineed to include other new material elsewhere and yet keep the book toroughly the same size. Given the title of the book I decided to remove thatsection. Instead I would recommend that interested readers look at Banister,Burman, Parker, Taylor, and Tindall (1994), Hayes (1997) and Smith (2008).These provide an introduction to a number of such methods and referencesfor those wishing to pursue them further.

Is psychology a science?The classic view of science is that it is conducted in a number of set stages.Firstly, the researcher identifies a hypothesis which he or she wishes to test.The term hypothesis is derived from the Greek prefix hypo, meaning less thanor below or not quite, and thesis, meaning theory. Thus a hypothesis is atentative statement which does not yet have the status of a theory. Forexample, I think that when people consume coffee in the evening they havepoorer sleep. Usually the hypothesis will have been derived from previouswork in the area or from some observations of the researcher. Popper (1972)makes the point that, as far as the process of science is concerned, the sourceof the hypothesis is, in fact, immaterial. While this is true, anyone assessingyour research would not look favourably upon it if it appeared to have orig-inated without any justification.

The next stage is to choose an appropriate method. Once the method ischosen, the researcher designs a particular way of conducting the methodand applies the method. The results of the research are then analysed and thehypothesis is either supported by the evidence, abandoned in the light of theevidence or modified to take account of any counter-evidence. This approachis described as the hypothetico-deductive approach and has been derived fromthe way that the natural sciences—such as physics—are considered to con-duct research.

The assertion that psychology is a science has been discussed at greatlength. Interested readers can pursue this more fully by referring to Valentine(1992). The case usually presented for its being a science is that it practises thehypothetico-deductive method and that this renders it a science. Popper(1974) argues that for a subject to be a science the hypotheses which it gener-ates should be capable of being falsified by the evidence. In other words, if




my hypothesis will remain intact regardless of the outcome of any piece ofresearch designed to evaluate it, then I am not practising science. Popper hasattacked both psychoanalysis and Marxism on these grounds as not beingscientific. Rather than explain the counter-arguments to Popper, I want toquestion whether use of the hypothetico-deductive approach defines a dis-cipline as a science. I will return to the Popperian approach in Chapter 10when I explain how we test hypotheses statistically.

Putnam (1979) points out that even in physics there are at least two otherways in which the science is conducted. The first is where the existing theorycannot explain a given phenomenon. Rather than scrap the theory,researchers look for the special conditions which could explain the phenom-enon. Putnam uses the example of the orbit of Uranus not conforming toNewton’s theory of gravity. The special condition was the existence ofanother planet—Neptune—which was distorting the orbit of Uranus.Researchers, having arrived at the hypothesis that another planet existed,proceeded to look for it. The second approach which is not hypothetico-deductive is where a theory exists but the predictions which can be derivedfrom it have not been fully explored. At this point mathematics has to beemployed to elucidate the predictions and only once this has been achievedcan hypotheses be tested.

The moral which psychologists can draw from Putnam’s argument is thatthere is more than one approach which is accepted as scientific, and that in itsattempts to be scientific, psychology need not simply follow one approach.Modelling is an example of how psychology also conducts research in theabsence of the hypothetico-deductive approach. Cognitive neuropsycholo-gists build models of human cognition from the results of their experimentswith humans and posit areas of the brain which might account for particularphenomena: for example, when an individual is found to have a specificdeficit in memory or recognition, such as prosopagnosia—the inability torecognise faces. Computer simulation is the extension of exploring a theorymathematically to generate and test hypotheses.

Ethical issues in psychological researchWhatever the research method you have chosen, there are certain principleswhich should guide how you treat the people you approach to take part inyour research, and in particular the participants who do take part in yourresearch. Also, there are principles which should govern how you behavetowards fellow psychologists.

Both the BPS (British Psychological Society, 2006) and the APA (Ameri-can Psychological Association, 2002) have written guidelines on how to con-duct ethical research and both are available via their websites. In addition, theBPS has produced specific guidelines for research via the Internet or Internet-mediated research (IMR) (British Psychological Society, 2007). Shaughnessy,Zechmeister, and Zechmeister (2009) outline the APA’s guidelines andinclude a commentary on them about their implications for researchers.

To emphasise the point that behaving ethically can have benefits aswell as obligations, I have summarised the issues under the headings of

12 Introduction



Obligations and then Benefits. I have further subdivided the obligations intothe stages of planning, conduct and reporting of the research. Many of thetopics covered are a matter of judgement so that a given decision about whatis and what is not ethical behaviour will depend on the context.

Obligations

Planning

As researchers, we should assess the risk/benefit ratio. In other words, weshould look to see whether any psychological risks, to which we are propos-ing to expose participants, are outweighed by the benefits which the researchcould show. Thus, if we were investigating a possible means of alleviatingpsychological suffering we might be willing to put our participants at morerisk than if we were trying to satisfy intellectual curiosity over a matter thathas no obvious benefit to people.

Linked to this is the notion of what constitutes a risk. The term ‘minimalrisk’ is used to describe the level of risk which a given participant might havein his or her normal life. Thus, if the research involved no more than thisminimum of risk it would be more likely to be considered ethically acceptablethan research which went beyond this minimum.

It is always good practice to be aware of what other researchers havedone in an area, before conducting a piece of research. This will preventresearch being conducted which is an unnecessary replication of previousresearch. In addition, it may reveal alternative techniques which would beless ethically questionable. It is also a good idea, particularly as a noviceresearcher, to seek advice from more experienced researchers. This will beeven more important if you are proposing to conduct research with peoplefrom a special group, such as those with a sensory impairment. This willalert you to ethical issues which are particular to such a group. In addition, itwill prevent you from making basic errors which will give your researcha less professional feel and which possibly make the participants lessco-operative.

What constitutes a risk worth taking will also depend on the researcher.An experienced researcher with a good track record is likely to show a greaterbenefit than a novice.

If risks are entailed which go beyond the minimum, then the researchersshould put safeguards in place, such as having counselling available.

Conduct

Work within your own level of competence. That is, if you are not clinicallytrained and you are trying to do research in such an area, then have a clinic-ally trained person on your team.

Approach potential participants with the recognition that they have aperfect right to refuse; approach them politely and accept rejection gracefully.Secondly, always treat your participants with respect. They have put them-selves out to take part in your research and you owe them the commoncourtesy of not treating them as research-fodder, to be rushed in when youneed them and out when you have finished with them. You may be bored stiff




by going through the same procedure many times but think how you feelwhen you are treated as though you are an object on a conveyor belt.

Participants may be anxious about their performance and see themselvesas being tested. If it is appropriate, reassure them that you will not be lookingat individual performances but at the performance of people in general.

Resist the temptation to comment on their performance while they aretaking part in the study; this can be a particular danger when there is morethan one researcher. I remember, with horror, working with a colleague whohad high investment in a particular outcome from the experiments on whichwe were working and who would loudly comment on participants who werenot performing in line with the hypothesis.

Obtain informed consent. In other words, where possible, obtain theagreement from each participant to taking part, with the full knowledge ofthe greatest possible risk that the research could entail. In some cases, theconsent may need to be obtained from a parent or guardian, or even someonewho is acting in loco parentis—acting in the role of parent—such as a teacher.

Obviously, there are situations in which it will be difficult, and counter-productive, to obtain such consent. For example, you may be doing an obser-vation in a natural setting. If the behaviour is taking place in a public place,then the research would be less ethically questionable than if you werehaving to utilise specialist equipment to obtain the data.

Although you should ideally obtain informed consent, do not reveal yourhypotheses beforehand to your participants: neither explicitly by telling themdirectly at the beginning nor implicitly by your behaviour during theexperiment. This may affect their behaviour in one of two ways. On the onehand, they may try to be kind to you and give you the results you predict. Onthe other hand, they may be determined not to behave in the way you pre-dict; this can be particularly true if you are investigating an aspect of humanbehaviour such as conformity.

If you are not using a cover story, it is enough to give a general descrip-tion of the area of the research, such as that it is an experiment on memory. Becareful that your own behaviour does not inadvertently signal the behaviouryou are expecting. Remember the story of the horse Clever Hans, whoappeared to be able to calculate mathematically, counting out the answer bypawing with his hoof. It was discovered that he was reacting to theunconscious signals which were being sent by his trainer (Pfungst, 1911/1965). One way around such a danger is to have the research conducted bysomeone who is unaware of the hypotheses or of the particular treatment agiven group have received and in this case is unaware of the expectedresponse—a blind condition.

Do not apply undue pressure on people to take part. This could be aparticular problem if the people you are studying are in some form of institu-tion, such as a prison or mental hospital. They should not get the impressionthat they will in some way be penalised if they do not take part in theresearch. On the other hand, neither should you offer unnecessarily largeinducements, such as disproportionate amounts of money. I have seen parti-cipants who were clearly only interested in the money on offer, who com-pleted a task in a totally artificial way just to get it over with and to obtain thereward.

Assure participants of confidentiality, that you will not reveal to others

14 Introduction



what you learn about your individual participants. If you need to follow uppeople at a later date, you may need to identify who provided you with whatdata. If this is the case, then you can use a code to identify people and then, ina separate place from the data, have your own way to translate from the codeto find who provided the particular data. In this way, if someone came across,say, a sensitive questionnaire, they would not be able to identify the personwhose responses were shown.

If you do not need to follow up your participants, then they can remainanonymous. For example, if you are conducting an opinion poll and arecollecting your information from participants you gather from outside asupermarket, then they can remain anonymous.

Make clear to participants that they have a right to withdraw at any timeduring the research. In addition, they have the right to say that you cannotuse any information that you have collected up to that point.

If you learn of something about a participant during the research which itcould be important for them to know, then you are obliged to inform them.For example, if while conducting research you found that a person appearedto suffer from colour blindness, then they should be told. Obviously youshould break such news gently. In addition, keep within your level of com-petence. In the previous example, recommend that they see an eye specialist.Do not make diagnoses in an area for which you are not trained.

There can be a particular issue over psychometric tests, such as personal-ity tests. Only a fully trained person should utilise these for diagnostic pur-poses. However, a researcher can use such tests as long as he or she does nottell others about the results of individual cases.

In research which involves more than one researcher there is collectiveresponsibility to ensure that the research is being conducted within ethicalguidelines. Thus, if you suspect that someone on the team may not be behav-ing ethically, it is your responsibility to bring him or her into line.

You should debrief participants. In other words, after they have takenpart you should discuss the research with them. You may not want to do this,in full, immediately, as you may not want others to learn about your fullintentions. However, under these circumstances you can offer to talk morefully once the data have been collected from all participants.

Reporting

Be honest about what you found. If you do make alterations to the data, suchas removing some participants’ scores, then explain what you have done andwhy.

Maintain confidentiality. If you are reporting only summary statistics,such as averages for a group, rather than individual details, then this willhelp to prevent individuals being identified. However, if you are workingwith special groups, such as those in a unique school or those with pro-digious memories, or even with individual case studies, then confidentialitymay be more difficult. Where feasible, false names or initials can improveconfidentiality. However, in some cases participants may need to be aware ofthe possibility of their being identified and at this point given the opportunityto veto publication.

Many obligations are to fellow psychologists.




If, after reporting the results of the research, you find that you have madeimportant errors you should make those who have access to the researchaware of your mistake. In the case of an article published in a journal you willneed to write to the editor.

Do not use other people’s work as though it were your own. In otherwords, avoid plagiarism. Similarly, if you have learned about anotherresearcher’s results before they have been published anywhere, report themonly if you have received permission from the researcher. Once published,they are in the public domain and can be freely discussed but must be cred-ited accordingly. You should also give due credit to all those who haveworked with you on the research. This may entail joint authorship if thecontribution has been sufficiently large. Alternatively, an acknowledgementmay be more appropriate. It can be a good idea at an early stage in theresearch to agree on who will be in the list of authors of any publications andthe order of the names, as, in psychology, the first named author is seen as thesenior author.

Once you have published your research and are not expecting to analysethe data further, you should be willing to share those data with other psycho-logists. They may wish to analyse them from another perspective.

Benefits

In addition to all the obligations, acting ethically can produce benefits for theresearch.

If you treat participants as fellow human beings whose opinions areimportant, then you are likely to receive greater co-operation. In addition, ifyou are as open as you can be, within the constraints of not divulging yourexpectations before participants have taken part in the research, then theresearch may have more meaning to them and this may prevent them fromsearching for some hidden motive behind it. In this way, their behaviour willbe less affected by a suspicion about what the research might be about, andthe results will be more valid.

If you have employed a cover story you can use the debriefing as anopportunity to disclose the true intentions behind the research, to find outhow convincing the cover story was and to discuss how participants feel.This is particularly important if you have required them to behave in away that they may feel worried about. For example, in Milgram’s experi-ments where participants thought that they were delivering electric shocksto another person, participants were given a long debriefing (Milgram,1974).

Another useful aspect of debriefing is that participants may reveal strat-egies which they employed to perform tasks, such as using a particularmnemonic technique in research into memory. Such information may help toexplain variation between participants in their results, as well as giving fur-ther insight into human behaviour in the area you are studying.

16 Introduction



SummaryThe purpose of psychological research is to advance knowledge abouthumans by describing, predicting and eventually allowing intervention tohelp people. Psychology can legitimately be seen as a science because itemploys rigorous methods in its research in order to avoid mere conjectureand to allow fellow psychologists to evaluate the research. However, incommon with the natural sciences, such as physics, psychologists employ arange of methods in their research. These vary in the amount of control theresearcher has over the situation and the degree to which the context relatesto people’s daily lives. Such research is often classified as being either quanti-tative—involving the collection of numerical data—or qualitative—to dowith the qualities of the situation.

Throughout the research process psychologists should bear in mind thatthey should behave ethically not only to their participants but also to theirfellow psychologists.

The next chapter outlines the preliminary stages of research.




Date post:	26-Aug-2020
Category:	Documents
Upload:	others
View:	0 times
Download:	0 times

First published 2010 - Amazon Web Servicestandfbis.s3.amazonaws.com/rt-media/pp/common/... · Part...

Documents